AI Data Annotator & Model Evaluator, Automated Marine Biology Annotation Pipeline (Honours-track Capstone)
Label StudioImageObject Detection
Led data labeling on a marine biology annotation pipeline, focusing on image-based fish and sea cucumber detection for model training. Curated, reviewed, and refined image labels using multiple industry-standard annotation platforms. Designed and improved labeling guidelines to address annotation challenges including occlusion and motion blur.
• Managed annotation quality using Label Studio, CVAT, and Roboflow.
• Involved in both two-class object detection and non-parametric classification tasks on underwater imagery.
• Applied and iteratively refined bounding box and segmentation annotations on 1024 px tile images.
• Supported model training and evaluation by ensuring high inter-annotator agreement and robust data curation.
Led data labeling on a marine biology annotation pipeline, focusing on image-based fish and sea cucumber detection for model training. Curated, reviewed, and refined image labels using multiple industry-standard annotation platforms. Designed and improved labeling guidelines to address annotation challenges including occlusion and motion blur. • Managed annotation quality using Label Studio, CVAT, and Roboflow. • Involved in both two-class object detection and non-parametric classification tasks on underwater imagery. • Applied and iteratively refined bounding box and segmentation annotations on 1024 px tile images. • Supported model training and evaluation by ensuring high inter-annotator agreement and robust data curation.
2026 - 2026