The Differences Between Semantic Segmentation and Image Processing

In terms of feature extraction, this really talks about one of the core elements of the technology that we're using. One commonality is the network does a really good job with extraction of screen features like individual laser spots and retinal tears. But similar to humans when there's a lot of noise in the image, which you and I perceive as just poor resolution, right? The image is not clear because there's media haze from any reason that often sort of decreases the confidence of the network of correctly identifying a feature. Some types of noise don't really affect network performance.