SIGNATURE WORK
CONFERENCE & EXHIBITION 2024

The application of interactive supervised learning in biomedical image segmentation promises significant enhancements in anomaly detection, disease diagnose, and treatment care. AI’s ability to discern complex patterns in imaging data and provide quantitative analyses has revolutionized radiation oncology and other medical fields. Emphasizing the integration of scribble-supervised learning and interactive learning, my research concludes new methodologies for interpreting medical images with greater accuracy, less manual labor, and less time consuming. The analysis for comparing the interactive supervised learning provides a systematic view for future implementation and research improvements and suggests future directions. The significant contributions of this research also include an explicit demo of the focused model in the research, CycleMix. Overall, through a rigorous comparative analysis, the work suggests that scribble-supervised methods, complemented by user interaction, not only refine segmentation outcomes but also pave the way for substantial improvements in diagnostic imaging.