Paper Presentation

Here are some slides I made to present this NIPS 2017 paper in our reading group:

Drawing inspiration from the popular VGG networks, the paper proposes using a deep convolutional neural network architecture with small convolutional kernels for segmentation of gliomas in MRI images. The authors discuss the relative advantages of using small kernels, and also explore the use of intensity normalization as a pre-processing step, which was unconventional in CNN-based segmentation methods. The proposed algorithm obtained the first position for the complete, the core, and the enhancing regions in Dice Similarity Coefficient metric in the Brain Tumor Segmentation Challenge 2013 database (BraTS 2013). ...

Given two shapes, $N$ samples are drawn from the edge elements of the shape. There are no specific constraints on these points - they can be either on the internal or the external contour of the object. Moreover, they also need not correspond to keypoints for the shape (such as maxima of curvature, inflection points, etc.), and although desired that the samples be uniform in spacing, this too is not a rigid criterion. ...

Introduction This paper presents a graph cut approach to the image segmentation task. Considering the image to be a directed graph with two nodes representing the source (object) and the sink (background), the authors propose a combinatorial optimization framework for image segmentation using $s/t$ graph cuts. This is the first global optimization object extraction technique that is extensible to beyond 2-D images. ...