[ML] Object Detection & Segmentation(2)

Object Detection

Detects individual object instances, but only gives box

Semantic Segmentation

모든 픽셀 이미지의 카테고리를 분류해주는 것을 말한다. 하지만 붙어있는 객체라면 하나의 객체로 판별을 하게 된다.

Instance Segmentation

Detect all objects in the image and identify the pixels that belong to each object

Mask R-CNN

What’s different in Mask R-CNN and Faster R-CNN?

Mask R-CNN has an additional branch for predicting segmentation masks on each Region of Interest (RoI) in a pixel-to pixel manner.

Faster R-CNN is not designed for pixel-to-pixel alignment between network inputs and outputs.