The field of computer vision and machine learning is moving towards more accurate and robust methods for tracking and analyzing human motion. Recent developments have focused on overcoming the limitations of traditional vision-based approaches, such as occlusion and instrumentation of the environment. New methods leveraging graph-based solutions, inertial measurement units, and ultra-wideband ranging have shown promise in achieving more accurate and reliable results. Additionally, the integration of process mining and machine learning has led to improved predictive capabilities in software development workflows and industrial manufacturing environments. Noteworthy papers include:
- Group Inertial Poser, which presents a novel approach for estimating body poses and global translation for multiple individuals using sparse wearable sensors and ultra-wideband ranging.
- GRAP-MOT, which introduces a new graph-weighted solution for person multi-camera multi-object tracking in highly congested spaces.