- Conducted in-depth research on Object Detection and Instance Segmentation, thoroughly analyzing and replicating key works by prominent authors, including [Zhang et al. 2022], [Wang et al. 2022], [YOLOv5 2022], [Liu et al. 2021], [Wang et al. 2021], and [Hurtik et al. 2020].
- Developed a complete machine learning pipeline, along with data processing and visualization tools for data generation and analysis.
- Achieved 90 mAP and 100 FPS (<10 ms) inference by creating lightweight models for detecting anomalies in noisy Railway and Catenary datasets.
- Built software to flag incorrect model detections and allow users to provide feedback through the system.
- Deployed the models using a TCP-based control software integrated into a PyQt Graphical User Interface (GUI).
더보기