BioCoach AI delivers real-time biomechanical feedback to prevent workout injuries at home

AI Fitness Coach Addresses Surge in Home Exercise Injuries

The pandemic changed how millions exercise, but it came with a painful cost. The U.S. Consumer Product Safety Commission reported a 48% rise in injuries related to at-home exercise during COVID-19, primarily due to poor form without professional guidance 1

1

. Researchers from Drexel University and Michigan State University have developed BioCoach AI to address this gap, creating an AI fitness coach that delivers real-time personalized feedback on exercise form through computer vision and biomechanical modeling 1

1

.

How BioCoach AI Uses Smartphone Camera to Analyze Body Mechanics

Unlike typical fitness apps that offer generic encouragement, BioCoach AI provides anatomy-specific corrections by processing video through two complementary information streams 2

2

. The first stream uses a 3D convolutional neural network to capture visual appearance and motion patterns, while the second estimates 3D skeletal movements, analyzing joint angles, range of motion, and exercise phases 1

1

. This AI-powered tool identifies joints most relevant to each exercise before providing feedback. For squats, it monitors hips, knees, and ankles; for push-ups, it focuses on shoulders, elbows, and wrists 1

1

.

Biomechanical Feedback That Explains Why Form Matters

Dr. Feng Liu, assistant professor in Drexel's College of Engineering and Computing who led the research, emphasized the system's unique approach: "Our goal was to build a system that does more than look at pixels and generate a generic comment. BioCoach exposes the model to 3D motion, joint angles and exercise-specific constraints, so the feedback can point to a concrete movement issue and explain why it matters" 1

1

. Instead of vague instructions like "lower your body more," the system provides precise biomechanical feedback such as "increase elbow flexion to 90 degrees at the bottom" with rationale like "increase hip/knee flexion to distribute load" 1

1

2

2

.

Training on Enhanced Exercise Video Dataset

The team built BioCoach using the Qualcomm Exercise Video Dataset, which includes hundreds of hours of exercise footage with time-stamped coaching feedback 1

1

. Researchers re-annotated this exercise video dataset with more detailed biomechanical targets and added explanations for guidance. The team contributed more than 2,400 notes to over 200 videos used to train and test the system 1

1

2

2

. These annotations helped prepare the large language model that provides coaching to users, with preserved time stamps enabling evaluation of both guidance quality and timing 1

1

.

BioCoach Outperforms Competition From Tech Giants

Researchers tested BioCoach against programs from NVIDIA, ByteDance, Alibaba, Salesforce, OpenAI, MIT, Shanghai Jiao Tong University, Chinese University of Hong Kong, Peking University, and Peng Cheng Laboratory in China 1

1

. The system outperformed Stream-VLM, a program from MIT and NVIDIA, on text quality and judged correctness, while showing improvements in anatomy-specific feedback accuracy 2

2

. The prototype was presented at the Conference on Computer Vision and Pattern Recognition, hosted by the Institute of Electrical and Electronics Engineers and the Computer Vision Foundation in June 1

1

.

What Sets BioCoach Apart From Existing Fitness Apps

Current fitness apps fall short in several ways. Apple Fitness+ and Mirror offer video-based workouts, but feedback is pre-recorded rather than dynamic 2

2

. Peloton's Movement-Tracking Camera counts reps and flags issues but requires dedicated equipment like Bike+, Tread+, or Row+, and doesn't explain the reasoning behind form corrections 2

2

. Google's Health Coach and Samsung Health analyze biometric signals but can't see users moving and therefore don't provide guidance for form 2

2

. BioCoach is the first system to combine skeletal reconstruction with a language model that explains the mechanical consequence of each correction 2

2

.

Future Development Could Prevent Injuries for Millions

The research team is working to add capabilities to estimate joint reaction forces and muscle activation patterns from video feeds alone 2

2

. Supported by the National Science Foundation, BioCoach has strong potential to become a smartphone app that offers personalized corrective measures for home exercise 2

2

. If deployed as a consumer application, it could make expert coaching accessible to anyone with a camera, helping to prevent injuries and support sustainable workout programs both indoors and outdoors 2

2

.