Exploiting Semantic Context for Anomaly Detection in Medical Images

• Published in: 2024 International Conference on Optimization Computing and Wireless Communication (ICOCWC) pp. 1 - 6
• Main Authors: Sony, Anubhav, Singh, Umesh Kumar, Kannagi, A
• Format: Conference Proceeding
• Language: English
• Published: IEEE 29-01-2024
• Subjects: abnormalities; Aggregates; Anomaly detection; approach; demonstrate; Image analysis; Performance evaluation; Prognostics and health management; proposed; Semantics; Wireless communication
• DOI: 10.1109/ICOCWC60930.2024.10470933

Anomaly detection in scientific photos is a swiftly evolving area of ultra-modern computer-aided prognosis and clinical image analysis. This paper proposes a unique method for anomaly detection in medical snapshots, exploiting semantic context-based statistics. Especially, a deep latest approach is adopted to make the most semantic context-based facts from a medical photo to offer better anomaly detection outcomes. A custom dataset containing 2000 scientific pix is hired to train the version. The experimental outcomes demonstrate the effectiveness of the proposed approach compared to baselines and work in phrases ultra-modern precision, do not forget, and F1-rating. Furthermore, the proposed technique is examined on multiple clinical datasets and achieves better overall performance in terms latest F1-score. With the proposed method, the overall performance of cutting-edge paradox detection devices improves substantially. Additionally, an evaluation of the modern-day performance of various datasets suggests that the proposed approach is suitable for exclusive present-day medical photos. Modern-day study findings reveal that the proposed technique is promising as a device for detecting anomalies in clinical photographs. This technique makes it easier to detect minor abnormalities in medical photos.