Mds-net: multi-scale depth stratification 3d object detection from monocular images

HIGHLIGHTS

• who: Zhouzhen Xie and colleagues from the Institute of Marine Electronic and Intelligent System, Ocean College, Zhejiang University, Zhoushan, China have published the paper: MDS-Net: Multi-Scale Depth Stratification 3D Object Detection from Monocular Images, in the Journal: Sensors 2022, 22, 6197. of /2022/
• what: The authors propose a depth-based stratification structure derived from the conventional Feature Pyramid Network (FPN) . The authors design a novel angle loss based on the consistency of IoU and angle to improve the prediction precision of boxes and the converge speed of training at the same time, which . . .
  

   

  

  If you want to have access to all the content you need to log in!

  Thanks :)

  Username or E-mail:

  If you don't have an account, you can create one here.