Content-Based Image Retrieval using Encoder based RGB and Texture Feature Fusion

Recent development of digital photography and the use social media using smartphones has boosted demand for image query by its visual semantics. Content-Based Image Retrieval (CBIR) is a well-identified research area in domain video data analysis. The major challenges CBIR system are (a) to derive semantics (b) find all similar images from repository. objective this paper precisely define hybrid feature vectors. In paper, encoded-based fusion proposed. CNN encoding features RGB channel fused with encoded texture LBP, CSLBP, LDP separately. retrieval performance different tested three public datasets i.e. Corel-lK, Caltech, 102flower. result shows class properties better retained features, helps enhance classification datasets. average precision Corel-lK 94.5% it 89.7% 88.7% f1-score 89.5% 88.5% improvement value implies proposed more stable deal imbalance problem.