@article{article_1500279, title={Efficient Image Retrieval in Fashion: Leveraging Clustering and Principal Component Analysis for Search Space Reduction}, journal={Erzincan University Journal of Science and Technology}, volume={17}, pages={638–649}, year={2024}, DOI={10.18185/erzifbed.1500279}, author={Köktürk Güzel, Başak Esin}, keywords={Fashion search, fashion image retrieval, clustering, principal component analysis, Resnet50, VGG19}, abstract={In this study, a novel approach using clustering techniques and Principal Component Analysis (PCA) for reducing the search space in fashion image retrieval systems is introduced. The study focuses on extracting high-dimensional feature vectors from images of clothing items and finding the same or the most similar product using these feature vectors, thereby narrowing the search space. The proposed method employs unsupervised learning algorithms to analyze high-dimensional fashion image feature vectors, grouping them into meaningful clusters. This enhances search efficiency and improves user experience. By reducing the dimensionality of feature vectors with PCA, computational costs are minimized. Experimental results demonstrate that the proposed method significantly accelerates computation time while maintaining an acceptable level of accuracy.}, number={3}, publisher={Erzincan Binali Yıldırım Üniversitesi}, organization={KOSGEB}