Study and Development of Content Based Image Retrieval (CBIR) Using Multiple Features on Android Platform

The need for more sophisticated and effective picture retrieval methods is rising because of the quick rise in image creation and sharing, particularly from mobile devices with high-resolution cameras. Accurate picture searching no longer requires traditional techniques that rely on metadata or simple pixel comparisons. Purpose: The goal of this work is to create a successful Content-Based Image Retrieval (CBIR) system that improves the speed and accuracy of obtaining pertinent photos from sizable databases. Design/Methodology/Approach: To transform picture content into comprehensive feature vectors, the suggested CBIR system combines several visual aspects, including color, texture, and shape, with updated image descriptors. Multi-feature analysis is then performed to extract similar images from these vectors. Findings: Compared to traditional techniques, the suggested system considerably increases retrieval accuracy and processing time, according to an evaluation on a dataset of 1,000 photos split into 11 categories. Research Limitations/Implications: The study only included 1,000 photos, and additional testing on bigger and more varied datasets could be necessary to confirm scalability and generalization. Practical Implications: The system improves user experience in picture retrieval applications like digital libraries, social media, and e-commerce by providing a scalable and reliable method for platforms and organizations managing large image databases. Originality/Value: By combining multi-feature picture analysis with optimized descriptors, this study offers a fresh approach to long-standing CBIR problems and opens the door for scalable, real-time image search in contemporary settings.