Image Segmentation
Product Overview
Semantic segmentation: Classify each pixel in the image (roads, buildings, sky, vegetation, human body, etc.).
Typical models: U-Net, DeepLab, SegFormer, etc.
Applications: Autonomous driving perception, medical image segmentation, remote sensing image analysis, industrial defect area localization.
Instance segmentation: not only categorizes, but also distinguishes different individuals within the same category (such as multiple people, multiple vehicles).
Typical models: Mask R-CNN, SOLO, YOLACT, etc.
Applications: Robot grasping, industrial sorting, security statistics, retail product recognition.
Panoramic segmentation: semantic+instance fusion, making instances for "objects" and semantic for "background areas".
Applications: autonomous driving, city perception, high-precision scene understanding.
Consult Us