AI-Powered Shooting Analysis System ‘BIOMECHLAB’ Revolutionizes College Basketball Training in Korea

Gachon University is breaking new ground in collegiate basketball training with the full integration of ‘BIOMECHLAB,’ an AI-powered shooting analysis system, into its physical education curriculum. Developed by Dr. Hyung-joo Lee, a sports science and physical education Ph.D., the Vision AI application promises a data-driven approach to skill development, marking a first for university-level basketball programs in South Korea.

The system, a collaborative project with coach Kim Seong-gyeom of the Han Kibum Hope Sharing Basketball School, utilizes Google’s MediaPipe technology to extract real-time joint coordinates and angles – of the knees, shoulders, elbows, and wrists – using only a smartphone camera. This eliminates the need for expensive wearable sensors, offering a cost-effective and accessible solution for analyzing shooting form. The system is capable of stable analysis at 30 frames per second, providing immediate feedback within a classroom setting.

‘BIOMECHLAB’ doesn’t simply register whether a shot goes in or misses. Instead, it dissects the shooting motion into four distinct phases – STANDBY, RAISING, SET, and RELEASE_AND_FINISH – employing a Finite State Machine (FSM) structure to quantitatively analyze the transitions between each stage. This allows students to move beyond simply knowing *if* their shot is successful, to understanding *why*, by examining patterns in joint angle changes, release timing, and consistency of movement.

The implementation at Gachon University is specifically geared towards sports science-based performance analysis education. Students will be able to compare and analyze their own shooting form based on the AI-generated data. Crucially, the system provides real-time feedback on elements critical to improvement, such as joint angle variability, release timing stability, and overall consistency of execution. This empowers students to self-correct and refine their technique.

Adding another layer of insight, ‘BIOMECHLAB’ displays average values reported in prior biomechanical research alongside the student’s individual data. For example, a 2022 study on three-point shooting (Irawan & Prastiwi, 2022) indicated an average knee flexion of approximately 103 degrees during the preparation phase, extending to around 167 degrees at release. The study similarly noted an elbow angle of roughly 135 degrees at the point of release, with shoulder angles typically ranging from 110 to 125 degrees. Wrist flexion at release is also identified as a key component for generating backspin.

Dr. Lee emphasizes that these values aren’t presented as definitive ideals, but rather as benchmarks for comparison. “Basketball technique training isn’t about rote repetition,” he explained. “It’s about understanding and recognizing your own movements.” He further stated, “Combining artificial intelligence (AI) and sports science data to create a data-driven, self-directed learning environment where students can analyze and modify their own movements is our goal.” He believes this initiative represents a pivotal shift in university physical education, moving away from lecture-based instruction towards a data-informed approach.

Students enrolled in the Gachon University course will install the ‘BIOMECHLAB’ application on their smartphones via the cyber campus announcement board. They will then utilize the app not only during class sessions but also during individual practice, facilitating continuous learning and reinforcement. Currently operating on the Android platform, the application is undergoing ongoing upgrades to address minor compatibility issues.

Looking ahead, Dr. Lee’s research and development plans include expanding ‘BIOMECHLAB’s’ capabilities beyond joint movement analysis. Future iterations will incorporate tracking of the basketball’s trajectory to analyze launch angles and incorporate visual alignment analysis during the shooting motion, ultimately creating a more precise and comprehensive basketball biomechanics analysis system.

The underlying Vision AI technology, based on MediaPipe, isn’t limited to basketball. It’s also being applied to analyze technique and provide correction in other sports, including running, golf, and tennis. This suggests a broader potential for advancements in sports science education and performance analysis across various disciplines.

The integration of ‘BIOMECHLAB’ into the Gachon University curriculum is significant as a real-world example of combining AI technology with sports science. It’s anticipated to pave the way for the development of AI-based basketball coaching platforms and further research into data-driven sports education.

The system’s ability to provide detailed, objective feedback on shooting mechanics has the potential to accelerate player development and enhance the overall quality of basketball training. For students at Gachon University, it represents a unique opportunity to learn and improve their skills using cutting-edge technology.

The next step for Dr. Lee and his team is to gather data from the initial implementation at Gachon University and Hongik University, with plans to conduct academic research on shooting accuracy changes, movement stability, and the effectiveness of learning transfer. This data will be crucial in refining the system and demonstrating its long-term impact on player development.

Archysport will continue to follow the development and implementation of ‘BIOMECHLAB’ and its potential to reshape the landscape of basketball training.