Deep semantic segmentation

Semantic segmentation refers to an understanding of an image at pixel level, that is, when we want to assign each pixel in the image an object class (a semantic label). It is a process to obtain coarse-to-fine inference. It achieves fine-grained inference by making dense predictions that infer labels for every pixel. Each pixel is assigned to a label with the class of its surrounding object/region. In this recipe, you will learn how to use a couple of deep learning (pretrained) models to perform semantic segmentation of images, using DeepLab V3+ and Caffe FCN.

Getting ready

First, download the pretrained model from deeplabv3_pascal_trainval_2018_01_04.tar.gz at
https://github.com/tensorflow/models...