Tutorial Series: Instance Segmentation on Vector's camera feed

Let's compare the performance difference between Instance Segmentation and Object Detection

This is part 4 of our series on how to develop a Machine Learning (ML) model to allow one Vector robot identify the contour of another Vector robot in its camera feed (aka instance segmentation). In Part 1, we explored the difference between object detection and instance segmentation. In Part 2, we examined how to efficiently construct a labelled dataset which will be used to build a ML model that can do instance segmentation. In Part 3, we trained a ML model to do instance segmentation, and used it to have Vector detect the contour of another Vector in its live camera feed.

This article concludes the series with a discussion of the performance tradeoffs between instance segmentation and object detection. As a recap to the two techniques, object detection aims to identify and label objects in a picture by constructing a rectangular bounding box around the object and then labeling the object. Instance segmentation classifies each pixel of the image and can thereby distinguish boundarie…