Creative Design Studio Interaction Body Language Recognition Video Dataset

In today's creative design industry, understanding the complex non-verbal communication among team members is crucial. However, existing body language recognition methods perform limitedly in dense interactive environments, struggling to accurately interpret subtle body movements and postures. This video dataset aims to tackle the technical challenges of body language analysis in creative discussions, enhancing recognition algorithms' accuracy through high-quality video samples. The dataset is collected using HD camera equipment in real design studio environments, recording natural and unobstructed team interaction scenarios. Quality control includes multiple rounds of stringent annotation and consistency checks, reviewed by experienced behavior analysis experts. The annotation team comprises professionals from the fields of computer vision and behavioral psychology, encompassing 30 members. Preprocessing steps involve denoising images, extracting keyframes, and annotating action sequences, with data ultimately stored in a structured MP4 format for easy researcher retrieval and analysis. The dataset boasts excellent data quality, with action annotation accuracy reaching 95% per frame, significantly surpassing similar datasets in consistency and completeness. It employs innovative multimodal annotation methods and data augmentation techniques to aid in developing more robust behavior analysis models. It addresses the industry issue of inadequate recognition accuracy, significantly enhancing team collaboration efficiency. Comparatively, this dataset features unique scene richness and contextual diversity, with its diversity and scarcity making it highly valuable for research. Also, the data's organization and structure provide high versatility and scalability in different research directions, supporting a wider range of application scenarios.