This is the official PyTorch implementation of the paper "MotionAGFormer: Enhancing 3D Human Pose Estimation With a Transformer-GCNFormer Network" (WACV 2024).

25 окт. 2023 г. · In this paper, we present a novel Attention-GCNFormer (AGFormer) block that divides the number of channels by using two parallel transformer and GCNFormer ...

In this paper, we introduce the MotionAGFormer, a novel transformer-graph hybrid architecture tailored for 3D human pose estimation. At its core, the ...

Official implementation of the paper "MotionAGFormer: Enhancing 3D Pose Estimation with a Transformer-GCNFormer Network" (WACV 2024).

Продолжительность: 9:13
Опубликовано: 6 дек. 2023 г.

This work proposes a novel attention-free spatiotemporal model for human motion understanding building upon recent advancements in state space models.

The current state-of-the-art on Human3.6M is MotionBERT (Finetune). See a full comparison of 50 papers with code.

From skeleton data in 2D space, MotionAGFormer architecture exploits both spatial and temporal features, along with other normalization and information fusion ...

In this paper, we present a novel Attention-GCNFormer (AGFormer) block that divides the number of channels by using two parallel transformer and GCNFormer ...

MotionAGFormer: Enhancing 3d human pose estimation with a transformer-gcnformer network. Mehraban, S., Adeli, V., & Taati, B. In Proceedings of the IEEE/CVF ...