Advanced Image Recognition with Mask R-CNN

We will cover the progression of deep learning technologies leading up to Mask R-CNN: Convolutional Neural Networks (CNN), Fast R-CNN, Faster R-CNN and Mask R-CNN. Originally, images were labeled with one object name per image, such as "cat". Then, bounding box rectangles were used to label one or more objects within an image. Now, with masking, the irregular outlines of the object are identified in the image, including overlapping instances of the same object. We will also provide a software demo.

Speaker: Aniket Rangrej and Ravi Ilango, ForHorn Systems