Image retrieval using wavelet transform and shape decomposition
Yen Do, Soo-Hyung Kim, In Seop Na, Dae Wook Kang, Jin Hyung Kim · 2013
In this paper, we propose an image retrieval system that uses both local and global shape features to retrieve the most similar images from the database. To obtain both features, some pre-processing steps, such as object segmentation using Minimum Error Thresholding and border extraction, are firstly carried out. After that, the Grid Based method is used to extract the global shape feature. The system divides the image into smaller areas and extracts local features by applying discrete wavelet transform and singular value decomposition. Finally, we compute the similarities between the global and local features of the query image and all the images in the database to give the most possible candidate matches as a result. The experimental results show the strengths and effectiveness of the proposed system.