CREBAS: Computer-Based REBA Evaluation System for Wood Manufacturers Using MediaPipe

Recently, musculoskeletal disorders (MSDs) caused by repetitive working postures in industrial sites have emerged as one of the biggest problems field health. The risk MSDs workers is quantitatively evaluated using NLE (NIOSH Lifting Equation), OWAS (Ovako Working-posture Analysis System), RULA (Rapid Upper Limb Assessment), REBA Entire Body etc. Methods used for posture analysis include vision-based and motion capture analysis. Vision-based a method where an expert with ergonomics knowledge watches manually analyzes recorded images. Although inexpensive, it takes lot time to analyze. In addition, analyst’s subjective opinions or mistakes may be reflected results, so somewhat unreliable. On other hand, can obtain more accurate consistent but its measurement equipment very expensive requires large space measurement. this paper, we propose computer-based automated system that evaluate, automatically consistently, order supplement shortcomings these existing methods. CREBA uses body detection learning model MediaPipe detect worker’s area images sets based on position face, detected face tracking model. set area, positions joints are tracked model, angles calculated joint inverse kinematics, then calculating degree load evaluation method. verify accuracy results system, compared them experts’ results. result experiment showed slight difference about 1.0 points between group those system. It expected ergonomic study will reduce workers’ labor intensity improve their safety efficiency.