Image retrieval using edge detection, RLBP, color moment method for YCbCr and HSV color space

An Image Retrieval (IR) system is used for accessing and retrieving the images from large image database. Content means the image features like color, texture and shape of the image. For Color feature, it is scaling and rotation invariant. It encrypts the color data they are a good component to use under changing lighting conditions. Three color moments are figured per channel (e.g. 6 minutes if the color model is YCBCR). For shape feature Edge Orientation Histogram (EOH), it speaks to the relative frequency of occurrence of 5 types of edges in every neighborhood a sub-picture or picture square. The sub-image is defined by partitioning the image space into 4×4 non-overlapping blocks So, thepartition of image definitely creates 16 equal-sized blocks regardless of the size of the original image. For texture feature, we used two features namely; Rotated local binary pattern (RLBP) and discrete wavelet transform (DWT). DWT provides an effective, scalable and intuitive representation of colors present in a area or picture. It is used to preserve the detailed contents of the images along with the reduction of the size of the featurevector and it describes the texture feature of an image. The implementation result based on precision and recall. The proposed precision is reached up to 88.37% and recall is decreased as compared to previous approach. For classification, Support Vector Machine (SVM) is used. It classifies the data with class labels. The distance is calculated with different similarity metrics like, Manhattan Distance (L1), Euclidean Distance (L2), Jaccard distance (JD), Hamming Distance (HD) and relative standard deviation (RSD).