Trying out Tensorflow object detection API with Infrared input

This video shows a demonstration of trying out the tensor flow object detection API with infrared input.

[@user-qi1mf5gj4r]

You provide a lot of great information in this video. I read your infrared Object Detection in medium. I needed to use an IR camera to find objects. How can I make an RGB image into an IR image to feed into the IR camera? Give me a solution to this. I've been working on this project for a long time..

[@wjkang619]

I am a bit confused about the system. Does this system use learned APIs with RGB images? Or did you learn something new for IR?

[@anujshah645]

It's the same lesrned API , I did no training here I just tweaked the way input is fed

[@jubinmathai4674]

I needed to detect humans using FLIR c2 IR camera images. Do you have a video explanation for the way infrared input has to be changed for RGB algorithm to read objects? I found your paper on "Deep learning approach for human action recognition in infrared images", was this the same algorithm used for human action recognition. Thank you

[@paulwolff2121]

He stacked the infrared image 3 times as can be read here: medium.com/@anuj_shah/object-detection-in-infrared-3f247dc3708

[@joshualekan8178]

will this same model work for non-infrared input? I'm trying to have the model be used for both infrared and daylight images. Thanks

[@anujshah645]

In principle it's should work, if not, you can fine tune a bit.

[@joshualekan8178]

@@anujshah645 thanks

[@vibekdutta6539]

Sir, how to shuffle the images and their labels correspondingly

[@anujshah645]

One way to do this can be as below: from sklearn.utils import shuffle #Shuffle the dataset x_shuff, y_shuff = shuffle(x,y, random_state=2) random state ensures it shuffles the same way every time, you can read about it.

[@vibekdutta6539]

Thanks sir, but I actually managed to get another way of doing it by model.fit(x,y, shuffle=true), thanks for showing it with the sklearn. Always appreciated your tutorials!

[@sokhibtukhtaev9693]

bro you have been away for many months? Hope to see your tutorials from time to time!

[@anujshah645]

Ya, I am not getting sufficient time after my work to make good detailed tutorials . I want to do this year, hope I get enough motivation to push myself to make some good tutorials

[@anujshah645]

Can you tell me one topic on which I should make a video this year

[@sokhibtukhtaev9693]

@@anujshah645 Personally, It would be great if you could proceed your tutorials on object detection with deep learning (darknet, tensorflow), and implementing some object tracking, recognition, counting using those model and weights in opencv integration. I have a feeling that opencv is getting serious about deep learning and opencv integration.

[@anujshah645]

@@sokhibtukhtaev9693 I will try my best(m not sure how well the video will be) to take out some time to prepare tutorials.

[@sokhibtukhtaev9693]

@@anujshah645 I am sure you'll do your best! Take your time and live your life to its fullest! Cheers, brother!

[@kartiktyagi898]

Arey yeh emiway kabse deep learning krne lgA?

[@anujshah645]

Emiway??