Versi bahasa Indo : • 1 Pengenalan Mask RCNN, MNC dan FCIS untuk...
** Support by following this channel:) **
This is the fourth part of my review on Mask R-CNN original paper. In this video, we discuss about how RoI pooling, RoI warping & RoI align work, continued with second modification in Mask RCNN which is decoupling class prediction and mask generation.
Again, I make this video under assumption that you are already know the basic of image processing (especially convolution operation using sliding window), neural network and CNN. If you're not, I would recommend you the see this Stanford lecture series on CNN. This is really great lecture : • CS231n Winter 2016: Lecture1: Introduction...
Original paper of Faster R-CNN : arxiv.org/abs/1703.06870
However, this is just my interpretation about the original paper. If you find any mistake, just tell me :)
Cheers,
Ardian Umam
Facebook : facebook.com/mamu.naidra
IG : instagram.com/ardianumam/
Linked-In : www.linkedin.com/in/ardian-umam/
コメント