Smart Mastitis Detection and Severity Classification using Thermal Imaging and AI

Abstract

Mastitis is a common inflammatory disorder affecting lactating individuals and livestock, leading to significant health complications and economic losses if not diagnosed at an early stage. Conventional diagnostic methods largely depend on manua l clinical examination and laboratory analysis, which are often invasive, time-consuming, and subject to inter-observer variability. To address these limitations, this paper presents a non-invasive, AI-assisted mastitis detection and severity classification system based on thermal imaging analysis. Thermal images of the affected region are acquired and uploaded for processing, where abnormal temperature distributions corresponding to inflammatory responses are identified. A trained artificial intelligence model analyzes the extracted thermal features to classify mastitis severity into mild, moderate, and severe categories, enabling early and objective assessment. In addition to diagnostic functionality, the proposed system incorporates a severity-aware therapeutic mechanism, wherein controlled localized heat therapy is automatically initiated to enhance blood circulation and alleviate inflammation in early-stage conditions. The integration of image-based diagnosis, automated severity grading, and decision-driven intervention establishes a closed-loop diagnostic-therapeutic framework. Experimental validation demonstrates reliable classification performance with rapid response time, reducing dependency on subjective clinical evaluation. Owing to its contactless operation, low cost, and scalable architecture, the proposed solution is well suited for deployment in both resource-rich clinical environments and medically underserved or remote regions. Future enhancements include mobile platform integration, adaptive therapy modulation, and large-scale clinical validation.