Image Retrieval Method Based on Image Feature Fusion and Discrete Cosine Transform

This paper presents a new content-based image retrieval (CBIR) method based on feature fusion. The deep features are extracted from object-centric and place-centric networks. discrete cosine transform (DCT) solves the strong correlation of reduces dimensions. shallow Quantized Uniform Local Binary Pattern (ULBP), hue-saturation-value (HSV) histogram, dual-tree complex wavelet (DTCWT). Singular value decomposition (SVD) is applied to reduce dimensions ULBP DTCWT features. experimental results tested Corel datasets Oxford building dataset show that proposed fusion can significantly improve performance compared using single type feature. slightly also tests variable factors affect performance, such as principal component analysis (PCA) instead DCT. DCT be used for dimensional reduction without losing too much performance.