10th International Congress on Information and Communication Technology

Authors - Andre Viviers, Bertram Haskins, Reinhardt A Botha
Abstract - Tracking gastropod chemical trails is time-consuming and error-prone. This paper argues that computer vision provides a viable alternative. Using selected image manipulation and segmentation techniques, an unlabeled dataset was generated. A simple K-Means clustering algorithm and manual labelling created a labelled dataset. Thereafter, a best-effort model was trained to detect gastropods within images using this dataset. Using the model, a prototype was created to locate gastropods in a video feed and draw trace lines based on their movement. Five evaluation runs serve to gauge the prototype’s effectiveness. Videos with varying properties from the original dataset were purposefully chosen for each run. The prototype’s trace lines were compared to the original dataset’s human-drawn pathways. The versatility of the prototype is demonstrated in the final evaluation by generating fine-grained trace lines post-processing. This enables the plot to be adjusted to different parameters based on the characteristics that the resulting plot should have. This research demonstrated that a gastropod tracking solution based on computer vision can alleviate human effort.