In the previous chapter, we studied different methods for feature extraction and image classification using Convolutional Neural Networks (CNNs) to detect objects in an image. Those methods work well in creating a bounding box around the target object. However, if our application requires a precise boundary, called an instance, around the object, we need to apply a different approach.
In this chapter, we will be focusing on object instance detection, which is also termed image segmentation. In the second part of the chapter, we will first see MOSSE tracker with OpenCV see various approaches to tracking objects in a sequence of image
Segmentation and tracking are, however, not quite interlinked problems, but they depend heavily on the previous approaches of feature extraction and object detection. The application's range is quite vast, including...