Computer vision
Computer Vision is a field of Artificial Intelligence (AI) that uses Deep Learning to extract information about the content in an image. There are a few common tasks in computer vision.
Image Classification
Classify the entire image into a specific set of classes, ex. dog, cat, etc. Or a simple binary case of damage and no damage. The main drawback is it gives one class for the whole image, so if the image contains both a dog and a cat it will present a problem. This can be avoided by converting it to a multi-label classification, where multiple labels can be predicted for one image.
Learn more
Object Detection
Detects objects within an image and draws a box around them. The benefit here is that multiple objects can be detected in the same image.
Learn more
Semantic Segmentation
Predicts classes on a pixel level, ex. background, cat, dog. The upside is that it provides a clear boundary between two objects of different classes. Ex. roads and sidewalks when used for driving, in a way that object detection does not. The downside is that semantic segmentation does not separate between objects of the same category that overlap.
Learn More
Instance Segmentation
Detects instances of objects and their boundaries. It is similar to semantic segmentation in that one gets contours at a pixel level, but solves the issue of what is a distinct object.
Learn More