YOLO: Non-max suppression (in Hindi)

[@raghavagrawal6263]

Wow, simple and perfect explanation in deep. Thank you sir.

[@KshitizVermaDL]

Thank you!

[@harshitchopra1551]

Very well explained sir thanks a lottt♥♥♥♥♥

[@PavanTripathi-rj7bd]

Thanks for the great explanation!

[@av1shek_ps]

From object localisation and object detection i have seen all your videos in this playlist in one day. Please add more videos quickly.

[@KshitizVermaDL]

Thanks for the comment and appreciation. Please understand that it takes a lot of time to create these videos. And since I don't want to compromise on the correctness, speed of upload will be low unfortunately.

[@rohitsuryawanshi7352]

Thank you for all the videos. All videos are very simple and easy to understand. As we have came long way till object detection, now may I please request for the Object Tracking using ROLO or DeepSort please.

[@KshitizVermaDL]

Thanks for the comment and the appreciation. I would like to take your advice, however, I don't want to get into an infinite loop covering only computer vision architectures. Next I will do DeepFace and FaceNet. After that I will switch to NLP and RNNs. Once I am done with them, I will switch to the math behind optimisers and regularization. Then I will do a second round of improving vision. Sorry for that!

[@rohitsuryawanshi7352]

@@KshitizVermaDL Alright, that sounds reasonable. May I pleases request for the practical implementation of what we learned so far? I feel, whatever we learned theoretically, if we implement the same then that would be really helpful to thoroughly understand the concepts and will feed confidence in us. As you know, at the end we have to implement in practice what we learn. Thank you.

[@KshitizVermaDL]

Thanks again for a valuable input. I, however, want to follow what Andrew Ng did. He first covered all the basic theory and then rolled out a specialization in Tensorflow. For anyone looking for implementation, before I reach there, I would like to recommend joining specializations offered by deeplearning.ai on Coursera

[@thetensordude]

PLEASE START DESCRIBING RECENT RESEARCH PAPERS

[@KshitizVermaDL]

I plan to do that in the second pass over CNN playlist.

[@manishswami877]

thank you so much sir

[@av1shek_ps]

Very nice, add more videos sir and try to complete it quickly and then please also add coding part means how to implement these thing

[@alishbafarooq5638]

Thank you so much

[@atharvapathak3239]

Great !

[@ankityadav-xd5ld]

Amazing

[@rohitshinde8478]

thank you

[@rohitsuryawanshi7352]

How will the bounding box will be generated for the test image? In this video, the red bounding box has been created by the 8th/second last grid cell, but on what basic it has created ? Also, the center of red bounding box which is in 8th grid cell manually given for the training image, however during the test image my car object is blank I mean it doesn't have neither center nor manually created bounding box. I hope you got both of my questions, else let me know I will clarify. Thanks.

[@KshitizVermaDL]

Will read all your questions properly and respond this weekend.

[@lakshay__bansal-of9jv]

Yes, I have the same doubt, Please reply if anyone knows..

[@shubhamprasad6910]

if the image has two separate objects from the same class then if we do the independent non-max suppression for each class. It will result in the detection of only one class. Please share your views.

[@KshitizVermaDL]

Each bounding box will detect only one object. It is possible for two different bounding boxes to detect an object of the same class. There are no strictly yes/no here.

[@harshitchopra1551]

@@KshitizVermaDL Can we run two independent bounding boxes here too?