Build a Deep Iris Detection Model using Python and Tensorflow | Keypoint Detection

[@NicholasRenotte]

Went back through this during my morning coding session, there were some definitely sketchy annotations in there. Will try to clean up the dataset before releasing!

[@keshavk2461]

hey Nicholas can you please share your best super mario bros with reinforcement training model with me please I really need it

[@ahmadimran1274]

Dataset hasn't been uploaded

[@harshitmalik7363]

can you plz share the dataset

[@harshitmalik7363]

dataset?

[@i8chickennuggets]

Hey man, i’ve recently started learning ml, been watching a ton of your videos. My mind is blown and i hope i can reach the level of expertise as you so i can follow along to your videos in real time instead of watching the same video 7000 times haha.

[@vikashchand.]

Nic! Killing it brother! Lesgo! 🔥🔥🔥

[@NicholasRenotte]

VIKKKK!! Hell yeah man!

[@harryfeng4199]

Duuuuuuude I was literally thinking of doing a project similar to this thank u so much cuz there r a lot of stuff here that I can reference.

[@NicholasRenotte]

Yessss go getem!

[@RocketGyan]

Hey nicholas, awesome as always. I just wanna give u one suggestion. when u install the dependencies, do explicitly mention their version number as when this video will get old, ppl will not struggle with dependencies. I have experienced that and just wanna give u this suggestion

[@NicholasRenotte]

Yeah true, I should probably lock them down. Good suggestion @RocketGyan!

[@reubenthomas1033]

Maybe you could also show us how to create a requirements.txt file, so we can use that just after we create a virtual environment

[@stavrosvogiatzis4016]

Excelent, love your projects !! Thanks Nicholas

[@DariusSingh-DS10]

Amazing work 🙌

[@MonikaSingh-nu5sg]

Amazing work again...Thanks a lot !!

[@merveozdas1193]

This video is that I expected😍

[@octupusprimeoctupusprime3374]

hey Nich, where can i get the eye tracking data folder?

[@proteus333]

Love the song!

[@harryfeng4199]

What tools/process would u recommend me use if I wanted to actually deploy such a project? Ie a url that the public can go to and give permission to use their webcam to do wtv, like iris detection

[@NicholasRenotte]

I’ve done it as a flask and react app. But I haven’t really got a best process yet, got a few weeks off dev, let me think about it and come up with something!

[@harryfeng4199]

@@NicholasRenotte thxxxxx

[@BerselliBianca]

is anyone having an error with the keypoints? I don't know why I get this error in the iris detection file: {function_node __wrapped__IteratorGetNext_output_types_2_device_/job:localhost/replica:0/task:0/device:CPU:0}} KeyError: 'keypoints'

[@Rahultanna]

yes, I am getting same error. Did your error solved? and if yes can you tell me how?

[@wheresthecode]

Thank you so much for all your videos, learning from your videos is so easy and enjoyable. I know you have done an Object Tracking video can you please please make one where we can count objects(people/cars) passing in/up and out/down a ROI (horizontal line for example) I have found a few examples but non are accurate. This will really help my grades

[@selinhaydaroglu3052]

I was wondering if you can train the model using the "Real Doppler RAD-DAR database" data, the radar data and the signal.

[@codybontecou]

What about tracking where you are looking on the screen?

[@raj4624]

Man you are sooo goood ... hands down.. Bow💎💎💎

[@mrudhhulavs765]

Can I know how long will it take to run 500epoch with this resnet152v2 model?

[@yogeshvasu5752]

does this model can be used for custom dataset

[@yashavanth3234]

Hiii... I am working on one project using gans which is 2d to 3d face generation I was totally confused how to train the modle and how can I predict that 😣 can you suggestions some.....☹️

[@harryfeng4199]

A shorter tut on the data collection part would be much appreciated, or maybe just some recommended resources that I can learn from. I find the labelme tool very interesting, already came up with multiple use cases that I can toy around with but I need to understand it better first

[@slideonme]

Man, he told like million times that you can check his Deep Face Detection video, where everything was explained step by step

[@darkcrawler8498]

Hi, I have a question. How to get the keypoint's label and display it in real time?

[@Cristina-TerezaTurcu]

Great video, I have one question, the code can also be used for iris recognition?

[@markmaara2012]

Hey Nick, great tutorial as always Question: you can also use the mediapipe face module to track your iris with refine_landmarks set to true, is creating your own model to track your iris better than the mediapipe model or what was your reason for not using the mediapipe model? Thanks

[@NicholasRenotte]

Mediapipe is great! Just wanted to show how you might go about building your own…but also how you might approach keypoint detection more generally. I’m working on a series of DL from scratch atm, hence this one! Keep using mediapipe, it’s awesome!

[@markmaara2012]

@@NicholasRenotte Thanks!

[@sushantp25]

Hey can anyone help me with the dataset? where to find it? Any sort of help would be really appreciated !

[@naifalharbi3121]

Can you Flatten the last layer rather than reshape?

[@sayedmohsen7714]

Is it okay to test this model on CASIA Iris dataset?

[@janhendrych1076]

Hi Nick, great video I'm just a little curious. Is there a particular reason why you didn't cache the data? Thanks

[@merveozdas1193]

if I do not haave gpu, is it not working well?

[@chitam727]

I'm new to Python. Do I need to watch "Getting Started with Python Deep Learning for Beginners" first?

[@udaybhandarkar]

where can i find the dataset ?

[@anahi_kr8235]

did you get tha data ? I am looking for it. please help me with data source

[@madhavgupta2002]

Where is the dataset?

[@sarveshsh591]

how did you convert point cordinates to keypoint

[@yolandasundari5356]

Can you make project about realtime ocr with python and tensorflow? Thankyou~

[@JatinPatel-oq7oe]

hello nich sir, i am getting error in following line cv2.imshow('object detection', cv2.resize(image_np_with_detections, (800, 600))) error: OpenCV(4.5.5) D:\a\opencv-python\opencv-python\opencv\modules\highgui\src\window.cpp:1268: error: (-2:Unspecified error) The function is not implemented. Rebuild the library with Windows, GTK+ 2.x or Cocoa support. If you are on Ubuntu or Debian, install libgtk2.0-dev and pkg-config, then re-run cmake or configure script in function 'cvShowImage'

[@arnavmehta3669]

Amazing, If I want to add recognition to it, do you have any idea how can I do that without spending a lot of resources (computational power) Thanks :D

[@ghaithalmasri3039]

Hey Nicholas, is there a way to reach out to you? I am working on a graduation project and would love your take on it

[@bhanudayaaturi2421]

How do i get data set?

[@reubenthomas1033]

Which monitor are you using ?

[@NicholasRenotte]

Samsung 32” T55

[@reubenthomas1033]

@@NicholasRenotte do you use a secondary monitor as well, or is a single 32" inch sufficient for you?

[@NicholasRenotte]

@@reubenthomas1033 have twin 32” T55s, probs overkill but I like having a ton of screen real estate. Helps a ton when training something off to the side or streaming. But just a bit of advice, don’t buy the T55 if you need to mount them, was an absolutely nightmare trying to find VESA adapters for them.

[@reubenthomas1033]

@@NicholasRenotte Got it! Would you recommend a 27" and a 32" setup?

[@NicholasRenotte]

@@reubenthomas1033 idk man I’ve got ocd and need symmetry but if you need a main a sub monitor that’s probably what I’d go with!

[@eraldpllaha7007]

Hello Nick! Can you help me please? I got an error from OOM, I restarted the kernel but again I get the same error. How can I fix this problem? Thanks!

[@pisitnakhonkhong2975]

Were you able to fix the probelm ? i was dealing with thesame issues but I found out that I had to decrease batch sizes from 16 all the way to 1.

[@ThatDirtySock]

Hey Nicholas, amazing video as always! All of your videos have greatly helped with effectively learning about the ins and outs of Data Science. I was wondering, do you have feedback, or even if you plan on doing a video, on the Math truly needed for a DS job? I understand what math you need to understand the underlying algorithms , but what about the actual job? Thank you!

[@Powercube7]

is this the fast thing u were talking aboot?

[@NicholasRenotte]

Fast thing? Actually yeah, you could target relevant points in your game!

[@whoami600]

nice bro

[@darkamax1]

Hi Nicholas, can you please do a video with YoloV5 and Deep Sort so that whatever we detect with the Yolo implementation we can track the same with deep sort. Thanks in advance.

[@venkatesh2788]

Please build recommendations system, with connection to the web app

[@twanashwany1642]

Hello Dear bro, how can I contact you?

[@aman1708]

Can you make a full-fledged Data- Science Playlist ? Please........

[@paufernandezpujol987]

Hey Nicholas, could you make a Rocket League RL bot tutorial, it would be amazing!

[@anfxf6513]

sir can you make a celebrity look like web app?

[@merveozdas1193]

zip=juup hoop

[@anfxf6513]

collecting data, cleaning data and create the model , after all create a web interface app

[@pablocaiza.]

if i want to detect multiple objects , what i need to change? thanks