A low-cost markerless motion-capture technique for early detection of illegal cricket bowling action

Abstract Cricket bowlers often practice new bowling varieties, which can lead to illegal actions such as straightening the elbows more than 15°. These illegal actions are reported by on-field umpires, forcing bowlers to undergo testing and action-correction training. The testing is typically conducted in sports-biomechanics laboratories using expensive marker-based motion-capture (MoCap) systems. This paradigm only addresses illegal bowling actions after they are reported in official matches, limiting its effectiveness in preventing such actions from developing in the first place. The traditional process often involves intrusive observations, as bowlers are aware of being monitored and may experience discomfort from the body-mounted markers. To address these limitations, the authors propose a low-cost markerless MoCap technique that can be used by coaches during practice sessions to observe the tendency toward illegal throwing and gradual deterioration of elbow angles. Our technique utilizes a camera setup that is easily deployable and can automatically track the joints of cricket bowlers using a particle-filter-based probabilistic approach. The data collected are then used to approximate elbow-extension angles, providing insights into the quality and legality of cricket bowling. The results were compared with visual marker-based observations with an average root mean square error of 0.1 pixels in joint positions. Moreover, our technique can also extract information related to other parameters of cricket bowling, including the horizontal crossing of the bowling arm, the approximation of ball speed and the ball release angle. By providing coaches with on-ground feedback on a bowler’s action, our technique can help to identify and correct potential issues before they escalate to the point of being reported in official matches.