The paper describes a new method to extract and cluster image features for effective still image databases. The feature vectors concerning color and texture are extracted using the multiresolution wavelet. Contrast to traditional image databases where feature vectors extracted from stored images are stored and used to match the feature vector of the input image for similarity retrieval, we use the Self-Organizing Maps neural network for clustering stored images. No feature vectors are stored in the databases, which saves storage space. A prototype image database is developed and some experiments are performed using it. The paper reports on the architecture and experimental results.