In this chapter, we will learn about various semantic segmentation techniques and train models for the same. Segmentation is a pixel-wise classification task. The ideas to solve segmentation problem is an extension to object detection problems. Segmentation is highly useful in applications such medical and satellite image understanding.
The following topics will be covered in the chapter:
- Learning the difference between semantic segmentation and instance segmentation
- Segmentation datasets and metrics
- Algorithms for semantic segmentation
- Application of segmentation to medical and satellite images
- Algorithms for instance segmentation