[Monocular Camera 3D Hand] 딥러닝 기반 3d-hand-pose-estimation

들어가기

새로 받은 과제. 스마트폰의 단안 카메라(Monocular Camera)로 손의 3D 좌표를 구하라!

 

기존에 널리 쓰이는 MediaPipe를 활용하는 방법, 딥러닝 기반으로 3D 좌표를 추정하는 방법 2가지를 알아보자.

 

 

딥러닝 모델 알아보기

 

NYU Hand, ICVL Hand, MSRA Hand 상위권 모델 3개

논문 및 코드 분석

 

• AWR

https://paperswithcode.com/paper/awr-adaptive-weighting-regression-for-3d-hand

Papers with Code - AWR: Adaptive Weighting Regression for 3D Hand Pose Estimation

• TriHorn-Net

https://paperswithcode.com/paper/trihorn-net-a-model-for-accurate-depth-based

Papers with Code - TriHorn-Net: A Model for Accurate Depth-Based 3D Hand Pose Estimation

• Virtual View Selection

https://paperswithcode.com/paper/efficient-virtual-view-selection-for-3d-hand

Papers with Code - Efficient Virtual View Selection for 3D Hand Pose Estimation