2017 Feature Pyramid Network for Object Detection (FPN) paper summary

[@nunorodrigues3195]

By far the clearest explanation, both logic and implementation wise, of an FPN.

[@joliver1981]

So I end up with 4 predictions in the end? Is that right? How should I use this output in practice?

[@MrSkifOK]

You help me a lot in deeper understanding! Thank you =)

[@TheSpecial43004]

instant subscribe, very clear in depth explanation, thank you

[@swedenontwowheels]

Great content! Very well done! I like how the concept is neatly and concisely presented without overwhelming information. Thank you very much for your effort in making this video.

[@zehra2334]

Keep making the videos! Very nice Hao Tsui

[@sourabhpatil9406]

Thanks for explaining! Will always be grateful. All the best.

[@MasterFX2000]

Very good explanation. Thank you!

[@annestephens553]

This was brilliantly explained - thank you so much

[@lovekesh88]

You earned a subscriber today. Thanks

[@nikunj.taneja]

Exactly what I was looking for, thanks! Keep making the videos man :)

[@ashishbhatnagar9590]

Very informative and helpful video. Thanks a lot for sharing.

[@originalandfunnyname8076]

Great explanation, thanks. You're videos really help a lot, I hope you'll do more of these summaries soon!

[@engintekin2530]

you did really good job explaining this paper in 17min. Thanks

[@Banefane]

Thank you very much for this clear and well thought explanation :D!

[@mudasirulhaque4873]

The finest explanation, cheers. Thanks for the content big help. :D

[@sujitha3335]

Thank you sooo much , pls keep posting more videos!!

[@mohammadyahya78]

Very helpful. Thank you.

[@andreestevam131]

Thank you for the clarification, very well explained!

[@mohaliyet]

Thank you man! 👏

[@walidaoun43]

Really Excellent explanation, Thanks :)

This is what im looking for :)) Thanks alot!!!

[@fpgamachine]

Excellent explanation, thanks. It was not clear to me how can I do to have a single softmax output for example.

[@shudiptosekharroy651]

this was really helpful! Please keep it up

[@zukofire6424]

Thank you! helps a a lot! :)

[@macaronminmin]

Thank you for your explanation !

[@ghassencherni7432]

Really good explanation, thank you!

[@minired4611]

Thank you so much. 謝謝

[@deoabhijit5935]

Thank you helped me a lot

[@galileo3431]

Great explanation! Highly underrated Channel!
To give some constructive criticism: Your explanation on the lateral 1x1 convolution may came a little short, considering the importance of this transformation. Your throat clearing makes it even harder to understand. But nonetheless, good work, keep going!

[@haotsui2720]

Thanks a lot for the constructive suggestions! Appreciate it

[@srikanthgn2969]

thank you so much

[@aisolutions834]

Perfect Hao, I've learned a lot from this high-quality video, I wanna watch all of your videos as being new to computer vision, Btw,
1. What new paper you wanna break down?
2. What do you think about Mnasnet?
3. Are there better object detection models than RetinaNet (fast while accurate)? as that also uses FPN.

[@aimerninelie2582]

Thanks a lot

[@adityajindal3738]

Great work buddy!!

[@goksuceylan8844]

thank you brooo

[@marx427]

Brilliant

[@moodguo4645]

The best ever seen

[@santoshkamble1290]

Thanks for this explanation!!!