Combining joint, limb, RGB, Taylor-video, optical-flow, and depth streams with two video backbones and a validation-tuned weighted ensemble reaches 73.213% top-1 accuracy on iMiGUE, the best MiGA challenge result to date.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
MM-Gesture: Towards Precise Micro-Gesture Recognition through Multimodal Fusion
Combining joint, limb, RGB, Taylor-video, optical-flow, and depth streams with two video backbones and a validation-tuned weighted ensemble reaches 73.213% top-1 accuracy on iMiGUE, the best MiGA challenge result to date.