Movyent: A Browser-Based Platform for Accessible CPU-Based Animal Behavioral Tracking

In neuroscience, quantitative behavioral analysis has traditionally relied on expensive commercial software or complex open-source tools, which often require a significant level of programming knowledge and GPU acceleration. To increase access to behavioral tracking technology, we developed a free browser-based platform called Movyent. Movyent was designed for a resource-limited computing environment, built using Python and Streamlit, and can be fully executed with a local CPU only. The core computer vision engine performs contour-based tracking using the Mixture of Gaussians 2 (MOG2) background removal algorithm. It alsoprovides a keypoint-based posture estimation function using the YOLOv8-Nano model optionally. The system calculates standard behavioral indicators such as velocity, thigmotaxis, and freezing, and ensures the integrity of scientific data by maintaining missing values when occlusion occurs. In addition, complex behaviors such as rearing and grooming are identified through explainable rule-based classifiers instead of opaque machine learning models. Movyent processes common video formats to extract accurate space and time data and generate paper-level visualization results such as kernel density heatmaps, trajectory plots,and zone occupancy charts. Researchers can easily export tracking coordinate data and aggregate behavioral indicators in a standardized format. Combining an intuitive graphical interface and a transparent computer vision technique, Movyent provides a powerful free alternative to conducting Open Field Test and variousbehavioral experiment paradigms and significantly lowers the technical barriers required to perform precise behavioral analysis.