Content-Based Radar Image Retrieval Using CNN Models

This study introduces a technique, for content based image retrieval (CBIR) specifically designed for radar image datasets, which are known for their complexity and high dimensionality. Achieving an precision (mAP) of 93.6% and an impressive retrieval accuracy of 94.8% the proposed system leverages advanced machine learning methods to enhance retrieval precision and efficiency. Additionally the system demonstrates a 35% reduction in retrieval time compared to state of the art approaches. A comprehensive evaluation was conducted, encompassing performance metrics, confusion matrices and detailed tabular analysis to validate these findings. For applications requiring accurate image retrieval these studies underscore the systems ability to handle radar image datasets with increased precision and speed thereby substantiating its effectiveness. The robustness and efficacy of the proposed method indicate its potential for application across disciplines where radar imaging is essential.