Meeting Abstract
Functional morphologists commonly utilize high-speed video to record movement and then use these recordings to measure aspects of biomechanical performance. Methods that rely upon measurements taken from individual video frames typically require extensive processing time, which in turn limits the amount of data that can be feasibly collected. Many kinematic studies have been restricted to small sample sizes that limit the application of statistical analyses. Although collaborative teams of researchers can process larger amounts of video in a given length of time, different members of the same research group will frequently extract different measurements from the same recording. Observer bias is therefore another serious issue that undermines the statistical integrity of kinematic studies. Automated methods can significantly reduce the time required to extract biomechanical measurements from video recordings. This not only supports the collection of larger datasets, and therefore a more rigorous application of statistical analyses, but also eliminates errors introduced through observer bias. Here we discuss an automated technique for extracting performance data from high-speed videos of fish feeding. This method collects measurements of velocity, acceleration, gape distance, jaw protrusion distance, hyoid depression distance, cranial elevation angle, and a ram-suction index estimate for individual video recordings of fish feeding strikes. These measurements are widely utilized by a number of labs investigating the functional morphology of fish feeding.
