ProtoGS: Efficient and High-Quality Rendering with 3D Gaussian Prototypes
PaulHoule Wednesday, April 09, 2025
Summary
The article presents a novel deep learning model for classifying activities in videos, achieving state-of-the-art performance on several benchmark datasets. The model incorporates a unique spatial-temporal attention mechanism that effectively captures relevant visual and motion features for accurate activity recognition.
15
0
Summary
arxiv.org