The Bold Blueprint Podcast
The Bold Blueprint Avideh Zakhor Every person has a unique potential waiting to be unleashed
Hey Amazon Prime members, why pay more for groceries when you can save big on thousands of items at Amazon Fresh? Shop Prime exclusive deals and save up to 50% on weekly grocery favorites. Plus save 10% on Amazon brands like our new brand Amazon Saver, 365 by Whole Foods Market, Aplenty and more. Come back for new deals rotating every week. Don't miss out on savings. Shop Prime exclusive deals at Amazon Fresh. Select varieties. How to have fun anytime, anywhere. Step one, go to Chambocasino.com. Chambocasino.com. Got it. Step two, collect your welcome bonus. Come to top a welcome bonus. Step three, play hundreds of casino-style games for free. That's a lot of games, all for free. Step four, unleash your excitement. Chambocasino has been delivering thrills for over a decade, so claim your free welcome bonus now, and live the Chambolife. Visit www.chambocasino.com. So welcome to EE225B. Title of the course is digital image processing. My name is Avida Zukor. I'll be handing out a hard copy of this information. She to everyone that's also on our class webpage. It turns out the course admin took the wrong information, so she's redoing it and bring it to the class in like five minutes. The class website is this. Oh, thank you. I just announced it. Hold on one second, let me make sure everything is cool. Ah, that's the right version. Thank you. So let me pass this along. Okay. Does everyone have a information sheet now? Okay, great. So I'll just go over the information here and then if there's any questions we'll get to those. So as I said, the class website is this. Check it out frequently for any announcements or homeworks or other things. I'd like to go through the information. I'd like to get a quick show of hands. How many of you are undergrads? One, two. And how many are grads? And how many are neither? Like visiting from somewhere else? Okay. Just you also? No, just you. Okay, were you visiting her? Right. I got your email. Okay, very good. And the undergrads, did you guys receive an email about the time change? Sometimes in the end of December moving the class to 9.32.11. Okay, and how did you figure out it's a different time? There was a sign outside this morning. Yeah. Okay, very good. And you never received that. And the grad students, you never received such an email. Okay. There was, there's been a mix-up about the time. We, there was a thought to change it. But then Ruzha Abachi decided to teach the biomedical and machine processing course. Also, Wednesdays and Fridays, I forgot overlapping with 9.32.11. So we just decided to keep 12.32 to slot, even though it's really not my favorite. And many people either ate, in which case they're sleepy afterwards or they're hungry, in which case all they can think about is food. So it's not the most optimal time slot, but I think we'll just stick to it for now. So the glass meets here. Wednesdays and Fridays 12.32. How many of you have taken class in this room before? Okay. So there's a couple of things I'd like to to mention. First of all, as you can see, the way this room works is there's, there's an operator there that, that takes a picture of me, whatever I write on the board, or if I just sit down and write on a piece of paper on the table, on the desk, it gets projected on these TVs and these TVs so you can, so that's how, that's how you figure out what's going on. The course gets videotaped and I believe, maybe I can ask the operator, does it, it doesn't have real-time broadcast, does it? Okay, so it's, it's webcast live and it's archived, but please don't abuse this by not showing up. It's not very fun to teach to a room of like one student, okay? I would never know if that student is super smart, I can just go really fast speed and nobody else would understand. So do show up and participate in the class. The other thing is there's a microphone to your right, okay? So you need, every time you ask a question, you need to pick up and use the microphone. The camera will not go your face, nobody would know who you are. So your privacy is, is maintained, but that way, when, when the course is recorded and I answer a question, somebody is watching the tape, let's say you yourself couldn't make it to the class, you're watching the tape at night, you know what the question was, okay? So as a courtesy to the fellow students who, who, who are watching the stuff offline, or for yourself, if towards the end of the semester you wanted to watch one of the lectures, it's very important that when you, when you ask a question, use the mic. It's difficult to remember at the beginning, but I promise you, it'll become a habit in like two or three lectures. And, and so the, the beauty of it is that you can, you can watch the lectures offline, or you can watch it real-time from your, from your office in Corey or, or wherever your, your office is. Now, I must warn you, based on my understanding of other students who've, who've used the real-time broadcast, is that it has, the video quality is not as high as if you, if you just watch it later when the course is finished and the, the, the file has been stored on a disk. And hopefully by the end of the semester you'll understand the technical reasons why that is, it turns on real-time broadcast, it has a lot more technical challenges than, than something that's been pre-recorded, finished and sitting on some hard, hard disk. So, so the class does give you the flexibility. There are, there might be a Wednesday or a Friday during the semester that I, I might not be able to make it to this class. And if I don't find the substitute, we might move the class to Monday, but I'll, I'll usually announce that way ahead of time. So, you know, if you can't switch to Monday, then you can always watch the video of the course. So, that's another advantage of this room. It gives us a bit of flexibility. How many people have a conflicts for 1232-2? One, two, three. Okay. I'll try to, at this point, I, I don't have anything planned. Yeah, I, I, this one doesn't, there's nothing that, that's eminent, but if it does, I'll give you plenty of advanced warning. So, going through the list here, my name is Avidazakor. My office is in 507, Corey Hoff. My phone number is extension three dash six, seven, seven, seven. My email is written here. Avz@ecs.berkeley.edu. Office hours are right after the class on Fridays from two to three. Teaching assistant Cindy Lu, she's sitting at the, why don't you stand up? Her office is, her research office is in three of seven Corey Hall and her, her extension is three, one, five, eight, seven. The department might give her an actual TA office. Have they told you what it is? Not yet, but we'll, I'll announce that later and it'll be on the webpage. Her office hours, TA office hours are, and this should have been on the sheet, but it's going to be on the web. So, make sure this information gets on the web, right? In Tuesdays and Thursdays at 3.30 to 5 p.m. So, there is no discussion or section or anything like that for this course, but Cindy will give you a lecture. But, her office hours have been, have been designed so that they're immediately before the day the lectures are given, or home works or lab assignments are due, so that any questions you might have on the homework or on the labs, she can help you out sort all these things out. The course assistant is, hey Amazon Prime members, why pay more for groceries when you can save big on thousands of items at Amazon Fresh? Shop prime exclusive deals and save up to 50% on weekly grocery favorites. Plus, save 10% on Amazon brands, like our new brand Amazon Saver, 365 by Whole Foods Market, Aplenty and more. Come back for new deals rotating every week. Don't miss out on savings. Shop prime exclusive deals at Amazon Fresh. Select varieties. It's time for today's Lucky Land horoscope with Victoria Cash. Life's gotten mundane, so shake up the daily routine and be adventurous with a trip to Lucky Land. You know what they say, your chance to win starts with a spin, so go to luckylandslots.com to play over a hundred social casino style games for free for your chance to redeem some serious prizes. Get lucky today at luckylandslots.com. No purchase necessary. BGW Group void were prohibited by law 18 plus terms of condition supply. Zita Alvarez, she was the lady who just walked in and handed in the handouts. Her office is 253 Corrie. Her extension is 34976 and her email is Rosita@ecs.breck2.edu. It turns out the lab assignments and the homework assignments will be posted on the web. But as a matter of policy, I don't provide the solutions on the web. So those will be handouts and you can either pick it up on class or if you don't pick it up from the class, you can go to Rosita's office. She's there from about 8 to 430 every day and you can pick up and also any old problem says, great the problem says that you didn't pick up, you can pick up from her office. Okay, before I move on to the text and stuff, are there any questions on the base administrative type stuff? Okay, textbooks. So there might be a little bit of confusion on that as well. So I have traditionally taught this course using this book on the two-dimensional signal image processing by Jay Lim. This is a fairly old book. I think it was done in the 80s if I remember correctly. It says 1990. That must be the second edition but I used it as a grad student in the in the mid 80s. Now we're going to use the course is kind of organized in two parts. The first one third to maybe one half, not quite sure, would be more on two dimensional signal processing, 2D filter design, a little bit of computer tomography, etc. And then the latter two-third or maybe one half will be on straight image processing. I think you need to know something about signal processing and the ways in which 2D signal processing is different from one dimensional signal processing before we can do image processing. So and I didn't, I really felt that this book is too old to address image processing. So this, another required book is this book digital image processing by Gonzales and Woods and that's straightforward image processing. So the second half of Jay Lim's book is on image processing but it's a bit outdated. So we won't go through that, rather we'll go through this. Now I understand that there might be too expensive for all of you to buy two books. So you're welcome to come up with any creative scheme to reduce the cost. If you only have budget to buy one book, I would recommend that you buy this rather than this. I will post just the problem sets out of this. The first maybe four problems says we'll be out of this thing. But again it'll probably be difficult, it'll be good for you to have this book so you can read the material so you can do the problems accordingly etc. How many of you have gone to ASUC to buy the book? Is this available? Okay did they have this? Yeah I just told the department to order this so it should be there shortly. So I'll wait till this book is in stock before I sign a substantial amount of homework or anything like that out of it. Maybe the first one is not due, well this Friday but maybe the following Friday or maybe later just to give enough time for this book to arrive as well. Okay then there's a host of other books that I've listed here that are kind of recommended. Number three is a handbook of image and video processing by Al Bovik. It's academic press 2000. It's really what it is, is lots of different chapters each written by a different person. There's always a bit of a danger in those kind of books. It's good in a sense that you're getting the different viewpoints about a topic like the image enhancement. You've got five chapters by five different people on five subspecialties of image enhancements and that aspect of it is good. It's bad a little bit because some holes could be there, some inconsistencies in notation come up, becomes a little bit more difficult to follow but nevertheless I have it in my office it's a worthwhile book to have. The other classic in the field is by Naturbali and Hasko. Barry Hasko, they were both at Bell Labs when they wrote it as you know there is no Bell Labs to speak of today. Barry Hasko moved to Apple, he's at Cupertino and I think Aruna Travali is a Georgia Tech but anyway their book is called Digital Pictures and it's a pretty good book. The other book that's a classic in the field is by Bill Pat who at the time he wrote it he was at USC now. Well he moved to Sun from there at Sun Microsis in so many years but he's somewhere in the Bay Area now and there's many editions of his book and I believe what it is two editions or three three editions already. That's also a classic and then if you just want to have a book on video processing and not image processing Murat Tech Help who is at the University of Rochester just published a book, well 95 was okay say it's not just it's only nine years digital video processing by Tarnas Hall. Other useful references, Dajan and Mersro, this is this is pretty old 1984 is at least 22 years ago. It's multi-dimensional digital signal processing. There's a bit of an overlap between that book and this book and Jay Lin's book but this is more more updated than Dajan and Mersro. Oppename and Shafer, that's the classic DSP book that we teach in EE123. Tom Wong has a book on 2D digital signal processing pretty old. It talks about a lot about filter design, something that we won't emphasize too much here but just touch upon. Tom Wong is a University of Illinois. Then you got Mitra and Extram and Mitra is UCSB, Santa Barbara, two-dimensional signal processing pretty old again. Gonzales and Wint's digital image processing. Andrew's and Hunt's pretty old 1977. Andrew's pretty old 78. Shriver's book is worthwhile having. He teaches at MIT, he's now retired of course but he deals with more electronic imaging systems. And Jane, fundamentals is really much passing. He was at UCB, he passed away actually a year or two after he wrote the book. So, outline of topics that we go over and we'll start with a little bit of overview of systems and signals and LTI systems and maybe just one lecture at most. Let me talk about image reconstruction from partial information that could be either from Fourier transfer magnitude or from Fourier transfer phase. The ways in which image reconstruction is different from one dimensional signal reconstruction. Talk a little bit about computer tomography which is actually a topic that's probably covered at a lot more depth and original botchies course because it's used a lot in medical imaging. Talk about 2D Fourier transform and Z transform, two-dimensional DFTs, FFTs, FIR and I say here IRR filter design but we probably won't touch IRR filter designs. It turns out that in 2D signal processing you don't use IRR filters too much simply because you can't prove their stability. Therefore, people just don't use them too much. Then we move on topic 4 onwards, image processing. Basics of it, perception, enhancement, restoration, super resolution, image and video analysis, classification segmentations, edge detection. Talk quite a bit about compression and coding for both still images and video. Talk a little bit about communicating image on videos that have been compressed over various networks, multimedia networking. We'll talk a little bit about rendering and half toning of images and videos. Talk a little bit about acquisition and then if there's time we'll talk about different applications. We almost never get to topic 14. Instead, different students who are interested in different applications choose a topic to do their projects and I'll talk about that a little bit more now. So there's homework support. Hey Amazon Prime members, why pay more for groceries when you can save big on thousands of items at Amazon Fresh? Shop prime exclusive deals and save up to 50% on weekly grocery favorites. Plus save 10% on Amazon brands like our new brand Amazon Saver, 365 by Whole Foods Market, Aplenty and more. Come back for new deals rotating every week. Don't miss out on savings. Shop prime exclusive deals at Amazon Fresh. Select varieties. If you get lonely climbing Mount McKinley so to entertain myself, I go to ChumbahCasino.com. At ChumbahCasino, I can play hundreds of online casino style games for free like online slots, bingo, slingo and more. Plus I go to daily login bonus. It's just too bad that up here, I don't have anyone to share my excitement with. With Chumbah Life. Anytime, anywhere. Play for free now at ChumbahCasino.com. P-T-W room. No purchases are avoided by law. See terms and conditions, 18 plus. Approximately once every week or once every two weeks, depending upon the synchronization between the lectures and the homeworks and the labs and other things. The homeworks can be either written assignments or there can be MATLAB assignments or just pure C programming. They will be graded, yeah they will be graded and they'll contribute 70% to the final grade. Yeah. I'm like pretty familiar with MATLAB but I would consider myself like kind of a novice or beginner and C and I was wondering if that's gonna be like a really good problem. I was gonna get to that in just a second to check people's backgrounds. Just hang on for just a sec. Let's see. So homeworks that are laid will not be accepted unless you've discussed it ahead of time with me. And then there's the remaining 30% of your grade and you can do one of two things. If you have a particular mini project in mind that can be done over a semester, you can do that and then write the term report, project report by the end of the semester. If you choose that option by end of February, you're supposed to hand in your two or three page proposal as to what you want to do. I don't want you to give me a proposal that when a project that requires two years for completion. You really have to specify what you think you can get done in the time duration of the course between now and May and say I'll show this, this and this and that's it. Or if you don't want to do a project, you can pick an area of interest and I can give out suggestions later on in the semester. I'll give you a sheet of possible topics you can choose from and areas. So you can pick an area and do a literature survey. Read the latest papers on it, kind of describe what you think the state of the art is, discuss the advantages and disadvantages, pluses and minuses of the existing schemes or approaches in the literature and then end your paper by, essentially that's like a critique of the papers you've read. You know, this paper was good but the method has these deficiencies, this paper is bad or good or this builds on top of this, etc. So critique of the papers and then end the paper by saying if you were going to do research in that area, what do you think are the open problems and how would you approach it? It doesn't mean you go ahead and do it and program and get generate results. It doesn't mean you do all of that but kind of a conceptual thinking of if you were going to do research in that area, what are the things that you would do. So I will, I will make a suggestion of the topics that you might be interested in in doing your research on and towards the beginning of February or a couple of weeks. There is a conference on image processing, international conference of image processing, the proceedings of that is online. As I'm sure you all know you have access to what's called IEEE Explorer, right? Does everybody know that? No? Yeah, on your computers on campus and maybe even at home if you've set it up properly, you can have access to the database of all the IEEE papers, international conference on image processing is an IEEE conference. I check to make sure that all the last four or five years of the conference in the conference was in our online. So that would be a good way of figuring out what goes on in image processing, what are the hot topics and figure out what area you want to do, either your project or your term paper. So if I go into any deeper, I like to kind of get a little survey of the people we have in the course. The class is small enough that we can kind of try to adapt it to your needs. Some of you might be, for example, already in bio engineering or some other field and want to be doing research in that. I like to know how many bioengineers we have, how many mechanical engineers, how many electrical engineers and kind of tell me in one sentence why you're taking the course and that one sentence could be, I thought it would be fun, that's good enough. It doesn't have to be a specific kind of technical reason. And while we do that also, let me just pass along this sheet that to get your names, email addresses, and undergrad or grad, and then year that you're in. I'll pass this around if you can, if you run out of this page, just turn it around or the other page, I had it done. So why don't, and also just tell me how much of C programming experience you have because it was a question I was just raised. And we start. Maybe use your mic also. My name is Kiran Shichan. I'm working on the entire 40-coder project in premium now, APGA4. So I chose to take this card and pay for the coder. And you're a grad student and you're building it on your chip, FPGA. FPGA. Okay. This is for your PhD. Master. It's a math project. Okay. Who's your advice? Young rabbi. Okay, very good. And how much C programming experience? My name is Vijay Yule. I first year grad student. In eeks. In eeks. Okay. I've done a project in speech processing, so I'm interested in DSP. I figured image processing would be like an interesting class. Good. And C programming I've had probably about like a year or a class in C programming. Okay. Good. Hi, my name is Galen Reeves. I'm also a first year in eeks. And I'm just interested in this class because I'm interested in image processing. And I guess I've had a decent amount of C programming experience. Okay, great. My name is Chris Anderson. I'm a first year applied science and technology grad student. And I work at the ALS and we deal with a lot of tons of images and I thought it'd be. What kind of images do they like biological samples and diffraction patterns, tons of different stuff and LLS stands for ALS advanced light source. And I don't have a whole lot of program experience in C. Have you ever done it? Yeah, I've written a couple basic things. I haven't had like a formal class. All right. Hi, my name is Howard. I'm currently working on a speech passing project at the XC and that's DSP. So I guess I'm really interested in this code signals concept and plus I've done some image processing stuff in the past. I think this would be really interesting. And I've had plenty of C program experience. So what year are you in? I'm the first year. First year. Yes. Oh, okay. So you're working with Nelson Morgan. Yeah. Okay, very good. I'm Sam Prandler. I'm visiting scholar from Switzerland. Officially, I'm a graduate non-student here. This is my official status. I'm specializing in signal processing. This is why I'm here. And I've done a lot of C programming. And you come from EPFL? No ETH Zurich. ETH Zurich. Okay, very good. My name is Jimmy and I'm a first year vision science student. And I'm currently working in a project on image compression. So that's what I think that's a good project to take. Do you work with Stan Klein? Bruno O'Shalson. Oh, okay. Very good. About two years. Okay. My name is Casey Lee. I'm a first year grad student in vision science. Like I haven't had much C programming experience at all. I had one undergrad course in it. Oh, that's one undergrad course on C programming. It's C and C plus plus plus. But I honestly say I don't remember anything from it. Okay. But I like that a lot. Hey Amazon Prime members. Why pay more for groceries when you can save big on thousands of items at Amazon Fresh. Shop Prime exclusive deals and save up to 50% on weekly grocery favorites. Plus save 10% on Amazon brands like our new brand Amazon Saver 365 by Whole Foods Market. A plenty and more. Come back for new deals rotating every week. Don't miss out on savings. Shop Prime exclusive deals at Amazon Fresh. Select varieties. Hello, it is Ryan and we could all use an extra bright spot in our day, couldn't we? Just to make up for things like sitting in traffic, doing the dishes, counting your steps, you know, all the mundane stuff. That is why I'm such a big fan of Chumba Casino. Chumba Casino has all your favorite social casino style games that you can play for free anytime anywhere with daily bonuses. So sign up now at Chumba Casino dot com. That's Chumba Casino dot com sponsored by Chumba Casino. No purchase necessary. VGW group for where prohibited by law 18 plus terms and conditions apply you a lot and I'm I'm an optics person. I work on a comic imaging of the retina and that's why I want to take this course because there's a lot of even processing of the retina retinol images. And you work with Stan Kline? No, Austin Roarda. Okay. My name's Robert Held and I'm a first year student in bioengineering and I'm taking this class because I'm an efficient science lab right now that there's a lot of stuff with new displays. So I thought it'd be good to have some image processing and I've done job but I've done any C or C++. Okay. You've done any mat lab? I've done a lot of mat lab. Okay. All right. Hi, my name is Karthikar. I'm a fourth year undergrad student in ekes and I've done a course on C. I'm just I'm really interested in image processing because I like the course the prerequisite to this course there's no processing. Oh, you took it. You were my class in the fall? Yeah, I was in your class. Oh, I said you're one of those who I was watched it online, right? Well, you wouldn't believe that we had 40 students the first day and by the last day of lectures is like five people but there was still 40 enrolled and 40 took the final and all of those but this attendance just went way down. So I have to come with a scheme to motivate you to all attend the class like maybe we'll have an attendance like five percent or ten percent of the grade is attendance. Last semester I had a conflict because if you remember... Oh, I changed it. Right. So that's why I couldn't make it to class once of the time. Right. Right. That's why I didn't want to change again because a lot of the students complain about the last minute change. Okay. Great. And your name was once again? Karthik. Karthik. Okay. Hi, my name is Gary Lee. I'm a second year grad student bioengineering but I was EECS as an undergrad and right now my interests lie in MRI type research but I haven't really taken any engineering classes in two years so I thought it'd be fun to do some signal processing again. Okay. And let's see. I've done a good amount of MATLAB programming but very little C. Okay. Hi, my name is Maine. I'm a fourth-year graduate student E. I'm taking this course in P. because that's the area I'm most interested in. Yeah, and I don't know what exactly signal process, image processing here I would do it but probably compression and some edge detection that kind of stuff. I only took one undergraduate course on C and yeah I'm very comfortable with MATLAB. Okay. Very good. My name is Mary, I'm a first-year grad student in EECS and I am taking this course just to learn more about DSP and I also have just had one undergrad course in C. Mm-hmm. Okay. I'm first-year EECS and I haven't really done much undergrad DSP in my undergrad so I'm kind of like exploring right now. So I'm doing DSP, the grad level in digital image processing. Just check it out. And I've done some MATLAB and some C, but I'm kind of rusty. It's been a while. Did you, you did the prelims already, right? Did you do the prelims? I'm gonna take it in next fall. Oh, okay. Okay. Okay. Very good. Hello, I'm a, my name is Sergey Timofiev. I'm a second-year grad student in electrical engineering. Mm-hmm. And right now I'm working on a array processing and that's there I'm interested in. Mm-hmm. So that's what I'm taking. Array processing as applied to wireless communication. And now I'm working on acoustic beamforming. Uh-huh. For what applications? For, like microphone arrays. Oh, okay. Just, just, okay. Mm-hmm. With professor Ahmad Baha'i. Okay. Mm-hmm. So, and I'm just, I'm interested in array processing. Gotcha. Gotcha. And I've had, I've written several programs in C. Okay. Very good. My name is Allen. I'm a second-year eke's grad student. Um, I've done, I didn't have pretty good experience with C programming. And I'm interested in, um, image processing and video communication in general. And I'm taking this course just to explore it okay? Okay. Very good. Hi. My name is Tvohal. I'm a second-year grad student at eke's as well. So I'm taking this course to, just to give myself up to date with image processing again. And I'm fairly comfortable with C. Okay. Very good. Hi. My name is Steve and second-year graduate student at EME. And I take the SPS of my minor. So that's why I'm here. Okay. And I've done a lot in my lab. And I think there were several courses in C programming in my undergraduate year. Okay. Hi. My name is Kenneth. Um, I'm a fourth-year undergrad eke's student. And, uh, I just kind of wanted to learn the two-dimensional versions of the Fourier transforms. And, um, I have just one undergrad course in C. And, uh, I think I'm pretty good at that lab. Okay. Very good. Okay. So, depending upon how things go, uh, we might or might not, uh, do much C programming. I think, uh, of the, of the six of the five or six lab assignments for our MATLAB and, uh, and maybe we'll do a simple, uh, C programming assignment, but, um, but not, not something terribly complicated. So, so even if you've just programmed, let's say, in four-term, which I'm sure none of you even know what it is, um, you could easily adapt to it and, uh, and do the, do the assignment. Uh, generally speaking, though, if you, if you're a researcher or a practitioner in image processing, MATLAB is good for prototyping, but it doesn't cut it when it comes to implementing systems. It's too slow. And at the end of the day, you have to convert it to, to C. So, one of the things I always see among my graduate students when they start is that they've done a lot of MATLAB, but then, when they go to grad school to do research and they feel that it's the first time they do the C programming and they really have to learn this. So, it'd be good to, even if you're not very good at C, be good to use this course as an excuse to, to ramp up and speed up, uh, your, your C programming skills. But, but we won't, we won't make it too hard because, uh, clearly it's not a prerequisite for this course. And we don't want that to become the focus of the course. Okay. Are there any, any questions, uh, regarding the administration of the course or the logistics or anything like that? Is there any, um, let me, let me do a few more question and answer things. To some extent, you kind of already answered it as we've been around, but is there any one topic or topics that most of you are more interested in anything else? I mean, clearly the field is big, we're not going to cover everything in it in this one semester. Um, is there like any burning requests to cover something to make sure we cover X? Yeah. Um, I'm kind of really interested in like biometrics, like detecting, uh, faces and stuff from a huge database and fingerprints and things like that. I think it's pretty cool. Okay. Any other requests or go ahead say a miscompression because I'm really interested in the different formats and how they've come about evolved. So, um, kind of like using like multiple cameras and, you know, using them together. Uh huh. Any other requests or favorites or anything like that? Okay. And the, um, the last thing I want to check is your background in DSP. So, how many of you have taken a DSP class at the level of, let's say, Oppenheim and Schafer? Oh, okay. Okay. And how many have taken it, let's say, just signals and systems, but not DSP? Okay. So were you undergrad ed? Uh huh. Okay. What textbook did you use when you took it? All right. Oppenheim and Schafer? Hey, Amazon Prime members. Why pay more for groceries when you can save big on thousands of items at Amazon Fresh? Shop prime exclusive deals and save up to 50% on weekly grocery favorites, plus save 10% on Amazon brands, like our new brand Amazon Saver, 365 by Whole Foods Market, Aplenty and more. Come back for new deals rotating every week. Don't miss out on savings. Shop prime exclusive deals at Amazon Fresh. Select varieties. I'm Victoria Cash. Thanks for calling the Lucky Land Hotline. If you feel like you do the same thing every day, press 1. If you're ready to have some serious fun for the chance to redeem some serious prizes, press 2. We heard you loud and clear. So go to luckylandslots.com right now and play over a hundred social casino style games for free. Get lucky today at luckylandslots.com. No purchase necessary. BGW Group, void were prohibited by law, 18 plus terms of condition supplied. With Bernard Levy. There were two other hands. I studied optics in undergrad and we used optics textbooks. But you never took like signals and systems. I guess from more optics point of view. Okay, you might want to watch the video lecture for E123. I thought it last semester. So if you find at least the first third of the course, things are a little bit foreign to you and you don't know the notation. Because the basic signals and systems class is a prerequisite for discourse. It doesn't have to be 123 but kind of at least 120 which is up on time on Wilski. And then 123 is up on a reshafer type level. So use that as a resource or consult with the book. Okay. There was one other hand. Okay. So you've taken signals and systems and that's good. That's good. Okay, any other logistics things we ought to take care of? Yeah. So would you recommend getting the digital image processing for MATLAB book? No, don't spend your money. That then becomes three textbooks and that will be $300 just to buy textbooks. So no, if the priorities should be one this and two the other one. If I do assign homeworks of the other one, I'll just make it available on the web so you don't have to have the book. I've looked at the book. It doesn't teach you the fundamentals. It just teaches you how MATLAB works essentially. The MATLAB image processing stuff. So if we do have the with MATLAB one with that, can that replace that book or should we get that book separately? Because on the books that I recommended the other book. It recommended the MATLAB book. The MATLAB book. Can you return that? Yeah, I okay. Yeah, return that and try to get this because I just I just ordered it. Well, not just on Friday of last week. I told the department to also have the bookstores ordered this. Any other how many of you have got let's say a textbook for this course already? Okay, so very few. Okay, and let me just quickly. One thing I didn't ask how many people are planning not to take this course for credit or just they're just listening for fun. One, two, okay. And let me just put a dot next to your name. What was your name again? You didn't write your name down even. Okay. How do you spell your last name? Okay, I got you. Okay, so it's it's audit. Okay. Any other questions, comments, anything else? All right, I usually don't get into too much technical stuff in the first lecture and try to kind of motivate a little bit the the as to why we're doing what we're doing and if you want to do that little 80 switch the point underneath the computer we get that the initial one that will show you the cells that will display the cell computer. Where is the switch? It's underneath that little table that's right next to you. Yeah. Take the point and play laptop to play dial. Ah, okay. Got you. And I'm just going to fire up. I put instead of bringing my laptop here. I just put some stuff online so we can we can just. Okay. So I don't have to carry any back laptops or connect the AV system. Oh, my God. Okay. Does anybody know which is preferable? Okay. Yeah, I did choose classic express. Yes, that's faster. So many steps just to get the thing going. Okay, very good. Okay. So I'll refer to that in a little bit. But for now, I'm going to just take our notes out from Okay. And just about just about all the stuff I show you during the class the scan version of it will be available on the class website, which is this address usually by the end of the day after the lecture. So you don't have to feel like you have to copy everything. Okay. So I'm going to talk about two things in today's lecture. One of them is basically differences between one and multi-dimensional signal processing. And then the second one is this PowerPoint that I just fired up is it has all the artwork in Gonzales and Woods. And I thought it would be good to just talk about, chapter one of it is just introduction to image processing, why study image processing and what are some sources of imagery. And so I thought it would be good to show you some of the discuss a little bit about sources of images, where they come from, what are the image modalities, how does one generate form an image or acquire an image, etc. But let me start with this first. So why are we studying image processing as a separate field? And actually one of my least favorite thing about this class is that this screen blocks my view of this quadrant of the class. So I just raised my chair hopefully. And then the only thing that's chopped is Howard's face. But I think I can see most of you know good enough. Usually I don't use the PC and then I just flatten the lap the screen. But today we're using it, so I'm going to do that. So here we go. So generally speaking multi-dimensional signal processing, you have, if you can zoom in just as much as you can, thank you. We have a lot more data than one dimensional signal processing. Let's just take a speech as an example. I think some of you are involved with speech research at ICSI. So the human voice is, I don't know, depending upon whether you're an opera singer or your normal person or male or female, but at most at 10,000 samples per second is enough to capture the quality of it. That's how you sample a speech, it's about 10,000 samples a second. And then if you compress it, you can do all kinds of speech compression technologies and squeeze it to as little as 20/2 tablespoons per second if they need arises. On the other hand, if you just think of an example of a multi-dimensional signal, it's television. A multi-dimensional, by the way, could be two, three, or four. So an image is a two-dimensional signal. It has an X and it has a Y. Then a television signal is a three-dimensional signal because you have X, Y, and then you have a time aspect to it. And then you can even think about a four-dimensional signal and how is that? What if you have a scene which is three-dimensional, X, Y, Z, and that scene is changing as a function of time, then you have four dimensions. And you can think about sampling or capturing your two-dimensional signal, the three-dimensional signal, and the four-dimensional signal. Now one of the things that's different about images and it is for audio signals, for example, one-dimensional signal is that, "Hey Amazon Prime members, why pay more for groceries when you can save big on thousands of items at Amazon Fresh?" Shop Prime exclusive deals and save up to 50% on weekly grocery favorites. Plus save 10% on Amazon brands, like our new brand Amazon Saver, 365 by Whole Foods Market, Aplenty, and more. Come back for new deals rotating every week. Don't miss out on savings. Shop Prime exclusive deals at Amazon Fresh. Select varieties. Hey, it's Ryan Seacrest. Life comes at you fast, which is why it's important to find some time to relax a little you time. Enter Chumba Casino. With no download required, you can jump on any time, anywhere for the chance to redeem some serious prizes. So treat yourself with Chumba Casino and play over a hundred online casino style games all for free. Go to Chumba Casino.com to collect your free welcome bonus. Sponsored by Chumba Casino. No purchase necessary. VGW Group. Void where prohibited by law. 18 plus terms and conditions apply. If you have an image with X and Y coordinates and each pixel has its own X and Y coordinates, the notion of causality goes away. On the other hand, when you have a time signal, let's say an audio signal, each sample is either before or after another one. So causality is a lot more alive and makes a lot more sense in one dimensional signal processing. Now, if if one of your dimensions of say in the television signal is time, you can still talk about causality. This frame came before or after the other one. Coming back to here, if you think about the television signal, as you all know, the standard in this country is NTSC, which some people jokingly call not twice the same color, but we won't talk about that. So how does NTSC work? And then of course, in Europe, it's pal and second, depending on whether you're French-speaking parts or maybe non-French-speaking parts, but second, I mean in France and pal everywhere else. But the basics are the same. So the NTSC, you've got 30 frames per second and each frame consists of two fields, even an odd field. So basically, you have even lines at time t0 and then odd lines at a slightly later time and then even lines and then odd lines, et cetera. And that's what you see. And roughly speaking, if you just approximate this thing 500 by 500 pixels per frame, and each frame having 30 frames a second, and each sample has R, G, and V values and R, G, and V, the three color components. Each one has eight bits, that's 24 bits per sample. You end up getting seven and a half megasamples per second, each one 24 bits. So the number of bits you have for uncompressed television signal is a lot higher than what you would get for audio signals. Of course, all of these numbers are off, not that we have HDTV, right? You have flat screens, then you have 1,000 pixels by 1,000 pixels, and you have 60 frames per second, and each pixel is 24 bits, so the numbers are huge. So automatically, when you deal with two-dimensional and three-dimensional signal processing, the volume of the data that you deal with is much larger than one-dimensional signal processing. For that reason alone, a major topic in multi-dimensional signal processing is image compression and video compression. You have to be able to reduce the size as well as you just overrun your computer. The other day I was doing the back of an envelope calculation, say, users generally, why is it that now still image photo sharing has become a lot more popular and acceptable than before? Well, you buy your digital camera, you put it on a disk, you plug into this into your PC at home, and they're all there. And even then, each picture is 1,000 pixels, it's not terrible, even after you compress it, it's not terribly easy to send all 50 pictures that you took to your dad and mom in Los Angeles, or whatever they live, they choose up a lot of space. Now think about video, if you took a, you know, 30 seconds of uncompressed video, roughly speaking, at 720 by 480 resolutions, roughly speaking, is 1 gigabyte. For half the people at home, that automatically choose up their disk. And let's say you want to edit it, so each time you edit it, a new version gets saved, that's another gigabyte. And so the storage requirements for video, for example, goes up real fast, real quick. So there's a fundamental need to compress these signals just to be able to handle them, put it on the PC, transfer it from the PC to somewhere else. And especially because at home, if you're connected with, you know, asymmetric lines like ADSL and cable, your uplink speed is much lower than your download, it's no more than 150 kilobits per second. So transferring data becomes quite a big challenge. As such, compression is a key issue in a lot of the multi-dimensional signal processing. The other thing is that the mathematics for multi-dimensional signal processing is not as complete as one-dimensional signal processing. And in fact, there are some parts of it that have shown for multi-dimensional signal processing, not to be able to do what some of the things that you have been able to do in one-dimensional signal processing. Examples of these are as follows. First of all, one-dimensional systems are described by differential equations. So if you have an IIR filter, you have a difference equation in continuous time, differential equation. In multi-dimensions, you have to deal with partial differential equations. Also, one of the things that makes things in two-dimensional signal processing much more difficult is that fundamental theorem of algebra does not hold in higher than one dimensions. And what is fundamental theorem of algebra? It's the idea that if you have a polynomial in one variable, you can factor it. If you have a polynomial of degree n, you can factor it into n simple factors of degree 1. That's well known. We all know how to do it. There's computer programs that does it for you. And what it results in is that if you've got an IIR filter, you can quickly find out by writing on the system function h of z as polynomial in numerator and a polynomial in denominator. By factoring the denominator, you can easily figure out what the poles of the system are. And therefore, you can easily check for things like stability. On the other hand, in two-dimensional signal processing, let's say for images, it's not that we haven't been able to figure out how to factor two-dimensional polynomials. But it's been shown that most two-dimensional polynomials are not factorable. And as a result, checking stability for two-dimensional IIR filters is extremely difficult. And it used to be one of the topics that I would cover maybe 10 years ago in this course. It's proven to be so useless that I don't even cover it. Basically, people in the literature have run aground with that. And almost all filtering that's done in digital image processing these days or multi-dimensional semesters is FIR filtering, which has no stability issues associated with that. And I'll talk about what stability means just as a reminder to most of you in just a few minutes. So bear with me. So that causes some major differences between 1D and multi-dimensional signal processing. And we talked about this third point, the IIR stability of filters. And I also talked about causality. The notion of causality really doesn't hold in two-dimensional signal processing. You can't say this pixel came before or after the other one was for time signals. You could easily say that. So let me just, before I get into the mathematics and start talking about notations and various other things, let me quickly go over some examples of images that one encounters in practice. A little bit of history. This is a, on the top, figure 1.1 sample of one of the earliest pictures that was produced in 1921 from a coded tape by a telegraph printer. This is an example of a picture that was transmitted across the Atlantic from London, New York, again 1929 or so. But really digital image processing took off around 60s. Ad JPL was a proportional laboratory in Pasadena, was one of the early places where image processing got started that followed by the name of Castle Man, did a lot of work there. And the images typically would come from the space. And you'll have the space missions. They took some pictures and they sent it down to us. So this is a first picture of the moon by a U.S. spacecraft, which was called Ranger 7. So as you can see, by now these pictures are quite routine. By the way, I don't know how many of you are following the stardust. Has anybody read about it in the newspapers recently? Anyone? Okay, you did. They made, they made available approximately one and a half million pictures. Now they're signing up volunteers. So it's basically, I have to know the details of it to say it properly. But basically they're trying to get samples of some particles from outer space, from the atmosphere. Hey Amazon Prime members, why pay more for groceries when you can save big on thousands of items at Amazon Fresh? Shop Prime exclusive deals and save up to 50% on weekly grocery favorites. Plus save 10% on Amazon brands, like our new brand Amazon Saver, 365 by Whole Foods Market, Aplenty and more. Come back for new deals rotating every week. Don't miss out on savings. Shop Prime exclusive deals at Amazon Fresh. Select varieties. Hello, it is Ryan and I was on a flight the other day playing one of my favorite social spin slot games on Chumbacassina.com. I looked over the person sitting next to me and you know what they were doing? They're also playing Chumbacassina. Everybody's loving having fun with it. Chumbacassina is home to hundreds of casino style games that you can play for free anytime, anywhere. So sign up now at Chumbacassina.com to claim your free welcome bonus at Chumbacassina.com and live the Chumbalines. Sponsored by Chumbacassina, no purchase necessary, VGW group, voidware prohibited by law, 18 plus terms and conditions apply. What planet was it? What was it? Some sort of a comet. So essentially the bottom line is we now have a canister that has about 45 or so good particles from an atmosphere of some comet and the idea is that if we understand the nature of those particles we can figure out the origin of life on earth for human beings. And one of the things that scientists have discovered is that life on earth didn't really start from earth itself. It was what was brought to earth by comets and other things, those particles that eventually got the thing going, that they had carbon dioxide or carbon or something that's needed for life to get started. But anyway the interesting thing is that meanwhile they also took one and a half million pictures when they were out there of this comet and they're now signing up volunteers to systematically go over these one and a half million pictures and detect the ones that have a particle. So actually when I read that a few days ago I thought that would be a great class project for this class. Maybe I'll look into it some more to see. The data is becoming available on the web but for volunteers to qualify they have to pass a test. Basically you have to look at an image and be able to detect it was a particle in there. And what strikes me is that why can we automate that? Why can we use the basic image processing tools? So I'll keep you posted. I have the name of the scientist and in fact the guy is at Berkeley Space Science Lab so it's not that far away. So we might be able to get access to that in which case that would be a really fun project to do during the semester in this course. You don't have to do each right, it'll be a team effort if you end up doing that. So here's examples of other images that are generated by other modalities. So on the upper left what you have is a bone scan, upper right it's a pet which is kind of a medical image, positron emission tomography. C is a sickness loop and D is a gamma radiation. X-ray image on upper left. Let's see a circuit board on the right. An angiogram on right here. This is an, okay let me use the mouse to point instead of doing upper right left and other things. Okay so this is a chest x-ray. This is an angiogram. This is a head CT. These are the eyes. This is a printed circuit board. It turns out a lot of image processing is used in inspection both in integrated circuit field in semiconductor industry. The biggest company in that field is KLA Tencore which is somewhere in San Jose. These guys have to inspect the masks using high resolution devices to make sure the masks for IC production don't have defects because if they do then everything is wasted. You've generated thousands of chips that don't function and the other field that's important is inspecting circuit boards to make sure that once you build a circuit board for electronic devices like DVD players this and that the other, they'll all function it. So here's another example of that. And this is an example of a sickness loop again. This is examples of ultraviolet imaging, normal corn and this is a smart corn. This comes from W. Davidson and Florida State University. Here's examples of light microscopy images. My favorite here is the middle one which is a cholesterol magnified cholesterol 40 times. This is a microprocessor as you would expect. Nickeloxide Tenfilm is used in semiconductor industry. Surface of an audio CD that's kind of interesting and here's an organic superconductor. So electron microscopes are another source of interesting imagery. Actually one of the exciting fields these days in the image processing is imaging, biological imaging of cells and molecules and stuff like that. So there's quite a bit of emphasis that's happening. Here's a Landsat images. These are images that have been captured by satellites and these are the different bands depending upon what wavelength you use and the same area using different wavelengths kind of looks different. So as you all know there's low flying satellites that take pictures, there's high flying satellites and there's all kinds of spying that goes on in the world. We get to know just about what every country is doing and now other countries have figured out how to send satellites up in the space. They get to find out what their neighboring countries are doing and how many of you are familiar with the Google Earth project? Okay, good. So if you haven't just go to Google Earth, download it and they've taken a bunch of images of Earth and you can pretty much figure out locate your home where you live and the neighborhood, et cetera. So there's a large large repository of database of images both from satellites, low flying airplanes and all kinds of other things. Also images are used in weather centers. So this is a multi-spectral image of Hurricane Andrew, which I believe was one of the major ones. So this is as you all know this is called the eye of the hurricane, what the center of it is. Here's an infrared satellite image of the Americas. As you know infrared looks at the heat sources. Infrared satellite images of the remaining populated part of the world. Here's some samples of manufactured goods often checked using digital image processing. We talked about that already. This is integrated. This is a circuit board. Pills, you can inspect it. Just make sure that everything, every one of these holes has a pill. You're filling up bottles. You want to make sure that the bottle is full and not empty. You are checking air bubbles and plastic cereal and also image of a intraocular implant. Actually, I don't know what that is. So you can use that in the manufacturing process to essentially figure out if things go wrong or if a product is being produced properly. Actually, there's a lot of image processing also used in agricultural business. Trying to make sure that the crops are healthy and figuring out what percentage of crops are, I don't know, corn, how much of it is this, how much of it is that. There's also a lot used in forestry. Last time I taught this because we had a student from forestry department. They take pictures of forests and using the heights of the trees, they try to determine how old the trees are and various other aspects of it. Here's some examples of images in the visual spectrum. Visual spectrum is the part that your eye sees. There's a visual part and there's infrared and there's ultraviolet. Infrared is beyond red where your eyes don't see and ultraviolet is beyond violet where your eyes don't see that. Here's a fingerprint on the upper left. This is a $100 bill and here's some pictures of license plates as the cars are passing by and I think in some transportation type applications. So as you know, there's camera and force intersections where if you pass a red light now they take your picture and they have the license plate number and you get a ticket that way. So that's where some of these technologies use. As a spaceborne radar image of mountains in south is Tibet. Do we have anybody from Tibet? From China? Okay. Does that look approximately right? Yeah, okay. Here's some MRI images. We had, I think, one or two people who were interested in magnetic resonance imaging. Here's an example of a bone. So why is magnetic resonance imaging important? Why is it different from x-ray? Well x-ray can only detect the bones. Magnetic resonance imaging can be used to look at soft tissue. So on your knee you can show the cartilage. You can see the ACL, MCL and don't ask me what does stand for ACL. They're all ligaments. MCL is internal to your knee. ACL is external and if you're not careful and you're teary, you'll be in excruciating pain probably for the rest of your life even though they can do surgeries. But you never really get fixed just from my own experience. I can tell you that. And so you could see with x-rays you could see the bones but you can't see all the fine details with x-rays. So MRI has, you know by the way, MRI is a fairly new technology. When I was leaving graduate school it was just getting adopted. 1988 or so is where you could finally walk into a center. I said I want to have an MRI and then you pop a thousand and they do it. And unfortunately the price hasn't come down over the last 20 years. It's still a thousand dollars per pop of MRI. The machines are allocated you know half hour slots and they need to generate I don't know 30 thousand dollars per machine per day to pay the cost of it over the years. On the right what you see is an MRI of a spine and that's extremely useful to figure out what your discs are doing. If the discs are protruded or touching the spinal canal, whether you have stenosis or whether you have arthritis or whether you have you know chipped discs, they're loose pieces hanging around your spinal canal. So these are soft tissues and what MRI does is it's imaging the hydrogen density and why hydrogen? Well because most of your body is made of water and so your discs are made of water. So MRI by by mapping the hydrogen density is essentially mapping the water density. That's how it distinguishes soft tissues from each other. That's how you figure out what the anatomy is doing. Here is some some NASA pictures of a of a Pulsar using gamma as as detectors x-ray detectors optical infrared and finally radio. It's all the same thing but you're looking at it at a different wavelength. So so that field by the way it's called multi-spectral imaging. That means you're looking at the same thing but at different spectral bands the images look different and by combining them you can make conclusions that you wouldn't be able to do by just looking at one of them. It's called multi-spectral or hyperspectral imaging. Here's some ultrasound pictures. Again there's a lot of examples from medical field. This is a baby right here. This is a baby upside down. Here's thyroids. Actually I don't know which one of these is a thyroid and this is some muscles. And here's an SEM image scanning electron micrograph. How many of you have worked in the micro lab and have used have played with SEMs or taken SEMs? Yeah okay so on the fourth floor of Corey Hall we have a micro lab where people fabricate you know structures or transistors or circuits or whatever. One of the main tools they use is scanning electron micrographs. After you build let's say a new kind of transistor what was the thing that Chairman Wu and Sujay and those guys invented FinFET. Now you want to take a three-dimensional picture of it and show it to the rest of the world. This is how this new transistor we invented looks like FinFET and you use SEM to do that. Or with lithography research you want to show you know three-dimensional depth on an integrated circuit. You shoot electrons and you collect the reflections of it and that's how you build the image. So the point here is that there's many ways and many modalities where you can take pictures of things. Here's some man-made images. These two are fractal images although they're not the oh sorry. Although they're not the best fractal images I saw. So we'll talk about fractal compression towards the end of the course but fractals are basically safe self-similar images and many years ago Barnsley over at Georgia Tech made fractal image processing popular and fractal compression popular. It turns out as a compression technique it's going to die off and it didn't really hold up to its promise. The classic case was when Barnsley took a sunflower field a picture of a sunflower field and he compressed it with fractal compression and he showed very good ratio and then somebody said okay now take that picture put an airplane up there now compress it and he could never do that because the airplane doesn't have self-similarity properties. The sunflower field did because the flowers are self-similar. So it's still a field that people do research you still see papers and I said people talking with you but it didn't really hold up to its promise. So these are two man-made images and these are these are images of objects that were generated using computer graphics. So you get a 3D model you'll random it from a particular view and then you can go around it and be with your multiple angles. So so images are sometimes even generated by us through a C program right that we just talked about. Okay so so here is kind of the outline of what at least Gonzales and Woods is we don't really follow these precisely but it's a good picture to take a look at anyway. So these are kind of steps that one takes in order to generate images. So the output of all these boxes are imagery. Image acquisition, image enhancement, image enhancement means what? You start with an image and you want to enhance something in it. It was too dark and you want to re-equalize for example the histogram so that you can see things better. That's an example of an image enhancement. Restoration something has gone wrong with the picture. The classic example, the Hubble telescope. It took pictures that were blurred. You couldn't figure out what's going on. That was after tens of millions of dollars spent to send the thing up there. So what do you do? You process it to restore it, undo a degradation that happened to it. Then there's color image processing, wavelets and multi-resolution processing that can be used for both processing images to do other things or for compression. And then there's compression itself using various techniques discrete porcelain transform, wavelets, this, that, and the other. Morphological processing which is yet another way of cleaning up images or processing images to get a new image. These are some examples of techniques that allows us to extract some attribute or some information about an image. Like you might do morphological processing in order to not only just generate a new image after you've done closure operations or opening or whatever, but you might want to extract something about it. How many objects were in the scene? Segmentation. Did you start with an image you divided into segment? That's a classic field. Many, many, many years of research has gone into that. In fact, David Forsyth and Jutandra Malick computer vision department here have done a lot of good work. If you look at the ERL research summary book, I think was it last year or the year before they had a picture of a zebra and they had done a nice segmentation of it. Then there's representation and description and finally there's object recognition. These are higher level computer vision stuff. We probably won't get a chance to touch upon this too much, but of course if you want to do your project on this, you're more than welcome to do so. I think that's about it. I'm going to stop now. I actually will not get into talking, but I was going to talk about two-dimensional sequences and delta functions and linear shift and variance systems and stuff like that. I won't do that today because we're out of time and I think it's much better to start it next time. Let me pause here. Are there any questions or comments? Okay, so I'll see you on Friday 12.32. Hey Amazon Prime members. Why pay more for groceries when you can save big on thousands of items at Amazon Fresh? Shop prime exclusive deals and save up to 50% on weekly grocery favorites plus save 10% on Amazon brands like our new brand Amazon Saver. 365 by Whole Foods Market, a plenty and more. Come back for new deals rotating every week. Don't miss out on savings. Shop prime exclusive deals at Amazon Fresh. Select varieties. How to have fun. Any time. Anywhere. Step one. Go to Chambocasino.com. Chambocasino.com. Got it. Step two. Collect your welcome bonus. Come to top of welcome bonus. Step three. Play hundreds of casino style games for free. That's a lot of games all for free. Step four. Unleash your excitement. Chambocasino has been delivering thrills for over a decade, so claim your free welcome bonus now and live the temple life. Visit chambocasino.com.