05. Classification: Summary

heading

Classification: Summary

Summary, Further Research, New Vocab Terms

This is the end of our section on classification and object detection.

Note that we did not spend too much time on the actual methods for building networks for classification and object detection, and as you will see further in the lesson, there is more focus on segmentation, especially on performance metrics and coding exercises.

The reason for that is that classification problems in 3D medical imaging can leverage a lot of techniques used for 2D image classification, and the course on AI for 2D medical image analysis, which is a part of this nanodegree, already provides an excellent deep dive into some of the approaches for classification and object detection problems. We talked about some of the differences between 2D and 3D classification problems such as 3D and 2.5D convolutions and hopefully, through our convolutions exercise, got you a feel of how these work and how you would code one yourself. But if you want to ground yourself better in applying CNNs for classification and object detection problems, I suggest going through the course on AI for 2D medical image analysis.

Further Resources