CHARLES RIVER ANALYTICS, INC. — Department of Defense SBIR Phase I: N143-126

Description

Selecting appropriate clothing systems is essential for the optimal performance of American Warfighters. However, no current assessment system quantitatively assesses and compares the thermal characteristics of clothing systems. Instead, scientists must manually assess the relative performance of different options using thermal cameras and software designed to detect hot and cold spots in static, fixed scenes, like architecture. This manual process is slow, labor-intensive, and prone to ambiguities.<br>We propose a Skeletal Alignment System for QUantitative Analysis of Thermal Clothing Hotspots (SASQUATCH). SASQUATCH is a software system that estimates the poses of human subjects in thermal imagery and computes and compares metrics about the thermal characteristics of their clothing, both overall and for specific body parts. The graphical user interface enables users to explore results in detail by focusing on particular body parts and specifying comparisons between different clothing systems. SASQUATCH is based on our state-of-the-art pose estimation and gesture understanding software, which detects and tracks humans in IR imagery and identifies individual body parts regardless of pose. This robust framework enables real-time analysis of clothing during natural movements and postures, beyond what can be seen in static poses, using a single thermal camera and standard computing hardware.