TemPose-TF-ASF: Two-Stage Bidirectional Stroke Context Fusion for Badminton Stroke Classification

TL;DR

This paper proposes TemPose-TF-ASF, a two-stage bidirectional stroke context fusion method that improves badminton stroke classification by modeling rich temporal dependencies, leading to significant accuracy gains.

Contribution

It introduces a novel two-stage training strategy and an adjacent-stroke fusion module that explicitly captures bidirectional temporal context for enhanced stroke recognition.

Findings

Achieves consistent accuracy and Macro-F1 improvements over baseline models.

Demonstrates strong transferability and generalization on large-scale dataset.

Integration of ASF yields notable performance gains in various models.

Abstract

Accurate badminton stroke prediction is crucial for fine-grained sports analysis and tactical decision support. However, existing methods struggle to model rich temporal context. This paper introduces TemPose-TF-ASF (Adjacent-Stroke Fusion), a context-aware extension of TemPose. It enhances stroke recognition by incorporating stroke-type information from both preceding and subsequent strokes. A two-stage training and inference strategy is adopted. Preliminary predictions from the baseline model are reused as estimated temporal context. These predictions guide the joint optimization of the ASF module and the classifier. By explicitly modeling bidirectional temporal stroke dependencies, the proposed method can be seamlessly integrated into existing state-of-the-art models. Experiments on a large-scale badminton match dataset show consistent improvements over the baseline and its variants in terms of Accuracy and Macro-F1. Moreover, integrating ASF into other advanced methods yields notable performance gains. These results demonstrate strong transferability and generalization capability.