Semantic Segmentation with U-Net

[Anbu]

Hi Sir,

@thearkamitra
@arosacastillo
@AmmarMohanna
@XpRienzo
@reinoudbosch
@chrismoroney39
@paulinpaloalto

We are unable to understand the context here , please help to clarify

This is a lot of outputs, instead of just giving a single class label or maybe a class label and coordinates needed to specify bounding box the neural network unit in this case, has to generate a whole matrix of labels. What’s the right neural network architecture to do that? Let’s start with the object recognition neural network architecture that you’re familiar with and let’s figure how to modify this in order to make this new network output,a whole matrix of class labels.

Does the above statement meaning, final result of the U-net would be whole matrix of class labels right sir?

[paulinpaloalto]

The point of semantic segmentation is that it gives a specific label to each point (pixel) in the input image. That’s what he means by a “whole matrix of class labels”. For every point in the image, you want a label that tells you whether that point is part of a car, pedestrian, driveable road surface, building … This is very powerful for applications like autonomous vehicles.