Efficient Content Based Image Retrieval Analysis of Distance Matrices
M. Vadivukarassi, G. JawaherlalNehru · Zenodo (CERN European Organization for Nuclear Research) · 2023
In this paper, we proposed a new method of feature extraction to improve the efficiency for retrieving the JPEG Compressed Images. We extract two DCT features, namely DC feature and AC feature, from the compressed image. Then we measure the image distance between the query image and the images in the database using these DCT features. Our retrieval system will give rank to the retrieved database images to define its similarity with the query image. Our proposed system does not need to full decoding, it only needs partial entropy decoding. Therefore, our proposed system takes less time for retrieving the images.