AiOS: All-in-One-Stage Expressive Human Pose and Shape Estimation

Ziwei Liu, Zhongang Cai, Yanjun Wang, Wenjia Wang, Lei Yang, Ailing Zeng, Chen Wei, Haiyi Mei, Wanqi Yin, Qingping Sun, Chi-Sing Leung

arXiv · PDF · Open in the Atlas

Code that ran

We have not lifted any functions out of this paper's repositories yet, so there is nothing we have run. If it links a repository, it is listed below.

Repositories linked to this paper

• ttxskk/AiOS — marked official in Papers with Code

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Encoder ℳ)"*' ℳ Decoder Figure 1. A comparison of existing methods in EHPS. (a) Top-down, multi-stage methods typically use detectors to detect humans, then use different networks to regress body parts on cropped images. (b) Top-down, one-stage methods use only one network for regression but still require detectors and rely on the cropped image. (c) Our all-in-one-stage pipeline, end-to-end human detection, and regression on full frame.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2403.17934")
get_code_for_paper("2403.17934")
have("2403.17934")

Connect an agent — have() is free.