You made a mistake. Selective search is not NMS ( Greedy Non-maximum Suppression), in fact Faster RCNN use NMS in one of the steps. Selective search is an algorithm that use segmentation to create region proposals (It uses NMS in one of the steps, but it's not NMS) . You can google selective search to see how it operates.
One small doubt Sir, you said the issue with stride is that with a stride of 2 we consider 6 out of the 7 pixels which leads to loss of data. But wont padding easily solve this issue? or was it found that padding is inefficient in this case?
Great Video, always really useful. One question Sir, when using Mask-R-CNN, do we always need to annotate the training data set to create and learn masks? or Mask R-CNN does it automatically?
Sir, I am working with ultrasound video for median nerve segmentation. I tried YOLO for detection, but didn't much accuracy, can you suggest me some model ? Thank you.
I move the video based on what it covers. In this case, it was covering a couple of lines that the viewer should be able to interpret. But I do understand your point. Thanks
Thank You very much my Guru.I never got a chance to be lectured by anyone else like you. Thanks a lot...
Thank You my dear computer vision mentor. Finally, wait is over and more exciting to watch object detection series.
Enjoy!
Thanks! sir, I hope the next video will also show how to use yolo in tensorflow in object detection.
Thank you. Yolo is not on my calendar yet, but will look into it.
Thanks so much for keeping uploading so many interesting and helpful videos!!! ☺
You made a mistake. Selective search is not NMS ( Greedy Non-maximum Suppression), in fact Faster RCNN use NMS in one of the steps.
Selective search is an algorithm that use segmentation to create region proposals (It uses NMS in one of the steps, but it's not NMS) . You can google selective search to see how it operates.
Most awaited video thank you Sir.!!
One small doubt Sir, you said the issue with stride is that with a stride of 2 we consider 6 out of the 7 pixels which leads to loss of data. But wont padding easily solve this issue? or was it found that padding is inefficient in this case?
Thanks Mr. Sreeni.
Please can you record videos about the DL framework "Caffe"
You are my best teacher.
Glad to hear that :)
Incredible explanation. Thanks😊
Thank you very much. Hopefully you can make videos about graph neural network
Amazing video. Thank you!!
great video. Thank you
Great video, pls add chapters to videos.
Great Video, always really useful. One question Sir, when using Mask-R-CNN, do we always need to annotate the training data set to create and learn masks? or Mask R-CNN does it automatically?
very well explained, thanks for your help buddy!
No problem 👍
Thank you so much !!!!!!! 🙏
very good explained :)
Thanks sir for all your videos, please I hope you can make videos of Faster R-CNN for objects detection.
Thank you, great video, i was needing exactly that, professor
Glad it was helpful!
can you do one on getting started with pytorch?
Sir, I am working with ultrasound video for median nerve segmentation. I tried YOLO for detection, but didn't much accuracy, can you suggest me some model ?
Thank you.
Thanks for the wonderful explanation! Kindly, can you minimize your live window a bit so we can see the whole slides :)
I move the video based on what it covers. In this case, it was covering a couple of lines that the viewer should be able to interpret. But I do understand your point. Thanks
Thanks sir for all your videos, i hope you make video of segmentation using conditional GAN or pix2pix GAN
Finally