# EgoCast: Forecasting Egocentric Human Pose in the Wild

Source: [EgoCast: Forecasting Egocentric Human Pose in the Wild](<https://www.sota2.com/research/paper/68fa75fe1e209131cea6acbf>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2025-10-26T03:45:34.560Z

Authors: Maria Escobar, Juanita Puentes, Cristhian Forigua, Jordi Pont-Tuset, Kevis-Kokitsi Maninis, Pablo Arbelaez

Published: 2024-12-03T23:11:06.000Z

## Abstract

Accurately estimating and forecasting human body pose is important for enhancing the user's sense of immersion in Augmented Reality. Addressing this need, our paper introduces EgoCast, a bimodal method for 3D human pose forecasting using egocentric videos and proprioceptive data. We study the task of human pose forecasting in a realistic setting, extending the boundaries of temporal forecasting in dynamic scenes and building on the current framework for current pose estimation in the wild. We introduce a current-frame estimation module that generates pseudo-groundtruth poses for inference, eliminating the need for past groundtruth poses typically required by current methods during forecasting. Our experimental results on the recent Ego-Exo4D and Aria Digital Twin datasets validate EgoCast for real-life motion estimation. On the Ego-Exo4D Body Pose 2024 Challenge, our method significantly outperforms the state-of-the-art approaches, laying the groundwork for future research in human pose estimation and forecasting in unscripted activities with egocentric inputs.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Head Rotation Prediction on CEN All Scenarios (test)](<https://www.sota2.com/research/sota/head-rotation-prediction-on-cen-all-scenarios-test.md>): Head Rotation Prediction; CEN All Scenarios (test); Head L1 Error: 0.0712.

- [Head Rotation Prediction on CEN High-U Scenarios (test)](<https://www.sota2.com/research/sota/head-rotation-prediction-on-cen-high-u-scenarios-test.md>): Head Rotation Prediction; CEN High-U Scenarios (test); Head L1 Error: 0.0716.

- [Trajectory Prediction on CEN All Scenarios (test)](<https://www.sota2.com/research/sota/trajectory-prediction-on-cen-all-scenarios-test.md>): Trajectory Prediction; CEN All Scenarios (test); ADE: 0.1092.

- [Trajectory Prediction on CEN High-U Scenarios (test)](<https://www.sota2.com/research/sota/trajectory-prediction-on-cen-high-u-scenarios-test.md>): Trajectory Prediction; CEN High-U Scenarios (test); ADE: 0.1198.

- [Uncertainty prediction on CEN dataset All Scenarios (test)](<https://www.sota2.com/research/sota/uncertainty-prediction-on-cen-dataset-all-scenarios-test.md>): Uncertainty prediction; CEN dataset All Scenarios (test); MAE: 0.1029.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2412.02903>)

[PDF](<https://arxiv.org/pdf/2412.02903.pdf>)
