Scriboora: Rethinking Human Pose Forecasting
This work addresses reproducibility and robustness challenges in human pose forecasting for applications like autonomous driving and human-robot interaction, though it is incremental in adapting existing methods.
The paper tackles reproducibility issues in human pose forecasting by evaluating existing algorithms, introducing a unified pipeline, and adapting speech models to improve state-of-the-art performance, with results showing substantial degradation from noisy poses and partial recovery via unsupervised finetuning.
Human pose forecasting predicts future poses based on past observations, and has many significant applications in areas such as action recognition, autonomous driving or human-robot interaction. This paper evaluates a wide range of pose forecasting algorithms in the task of absolute pose forecasting, revealing many reproducibility issues, and provides a unified training and evaluation pipeline. After drawing a high-level analogy to the task of speech understanding, it is shown that recent speech models can be efficiently adapted to the task of pose forecasting, and improve current state-of-the-art performance. At last the robustness of the models is evaluated, using noisy joint coordinates obtained from a pose estimator model, to reflect a realistic type of noise, which is more close to real-world applications. For this a new dataset variation is introduced, and it is shown that estimated poses result in a substantial performance degradation, and how much of it can be recovered again by unsupervised finetuning.