October 2017
Beginner to intermediate
270 pages
7h
English
Segmentation adds an additional layer of complexity to the mix because the model has to locate the elements in an image and mark the exact shapes of all the located objects, as in the following illustration:

One of the most common approaches for this task is to implement sequential downsampling and upsampling operations, recovering a high-resolution image with only a certain number of possible outcomes per pixel that mark the class number for that element.
Read now
Unlock full access