Abstract—Over the past decade, image retrieval has become the most important research subject in computer vision, attracting significant attention from researchers. The main challenge in image retrieval is significant image depiction and the extraction of large features. Hence, this research proposes an innovative method of cooperative learning between the actual Adam norm approach and Adaptive Moment Estimation with Gradient Correction (AMSGrad) within a Convolutional Neural Network (Adaptive Adam-CNN) for image retrieval. The proposed approach improves the discriminability of learned features and maintains high precision even for larger data. This research primarily employs the three standard datasets, namely Corel 1K, Corel 5K, and Corel 10K, to evaluate the effectiveness of the introduced approach. Next, a preprocessing technique, such as min-max normalization, is employed to scale the input data uniformly for improved image retrieval. The experimental findings demonstrate that proposed Adaptive Adam-CNN approach achieves higher accuracy of 99.63%, 98.37%, and 95.28% on Corel 1K, Corel 5K, and Corel 10K datasets, respectively, compared to the Content-Based Image Retrieval with Accuracy Noise Reduction (CBIR-ANR) and the Extended version of Local Neighborhood Difference Patterns (ELNDP).
 
Keywords—Adam norm, adaptive moment estimation with gradient correction, convolutional neural network, cooperative method, image retrieval, min-max normalization


Copyright © 2026 by the authors. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0).