Understanding the behavior of ants, their interactions and responses to various environ- mental elements offer a unique perspective on social structures. The ability to track ants movements accurately within a controlled laboratory environment is a valuable tool for biologists. This master's thesis explores ant tracking, a multiple object tracking problem, using advanced computer vision techniques. One primary challenge is the scarcity of suitable data with accurate annotations. This work addresses this issue by collecting new raw data and developing annotation tools. A YOLOv8n detector, with a validation mAP@50-95 of 85.5%, is trained and integrated into the tracking models. During testing, the detection performance decreases, with a mAP@50-95 of 15%. This significant drop, despite its notably low value, played a crucial role in enhancing the overall tracking accuracy. A BoT appearance descriptor for re-identification, achieving a validation Rank-1 accuracy of 74%, is trained and integrated into the tracking models. Subsequent testing within a crowded tracking environment, identifies a bad performance, leading to its exclusion from the final tracker. However, the analysis highlights the appearance model's potential for future investigation. The results, obtained using an OC-SORT tracker, establish a baseline for future research, achieving a 49% improvement in HOTA for the testing set. In conclusion, this master's thesis lays the foundation for future research by preparing data, identifying key components, and establishing an initial baseline.