Semantic Image Segmentation with Deep Convolutional Nets and Fully Connected CRFs

Deep Convolutional Neural Networks (DCNNs) have recently shown state of the art performance in high level vision tasks, such as image classification and object detection. This work brings together methods from DCNNs and probabilistic graphical models for addressing the task of pixel-level classifica... View More