错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Semantic Segmentation and Scene Understanding Using RGBD Data

  • Supratim Paul,
  • Abhijit Boruah,
  • Plaboni Mukherjee,
  • Bishal Basfore,
  • Debajit Sarma,
  • Manob Jyoti Saikia

摘要

Semantic segmentation is a crucial task in computer vision, aimed at classifying each pixel in an image into predefined categories. Traditional RGB-based semantic segmentation struggles with accuracy in complex, cluttered environments where objects overlap or vary in distance. This work addresses the challenges using a Microsoft Kinect sensor by performing segmentation on multiple objects in RGBD images, subsequently analyzing their spatial and depth information to determine the interrelations between different objects such as cans, boxes, and cups. This approach aims to enhance overall scene comprehension and object interaction analysis.