Adaptive color reduction

Abstract

The paper proposes an algorithm for reducing the number of colors in an image. The proposed adaptive color reduction (ACR) technique achieves color reduction using a tree clustering procedure. In each node of the tree, a self-organized neural network classifier (NNC) is used which is fed by image color values and additional local spatial features. The NNC consists of a principal component analyzer (PCA) and a Kohonen self-organized feature map (SOFM) neural network (NN). The output neurons of the NNC define the color classes for each node. The final image not only has the dominant image colors, but its texture also approaches the image local characteristics used. Using the adaptive procedure and different local features for each level of the tree, the initial color classes can be split even more. For better classification, split and merging conditions are used in order to define whether color classes must be split or merged. To speed up the entire algorithm and reduce memory requirements, a fractal scanning subsampling technique is used. The method is independent of the color scheme, it is applicable to any type of color images, and it can be easily modified to accommodate any type of spatial features and any type of tree structure. Several experimental and comparative results, exhibiting the performance of the proposed technique, are presented.

Multiresolution Histogram Analysis for Color Reduction

“…The most known methods to build a colormap with an a priori fixed number of colors are the median cut algorithm [1], which is based on the popularity method suggested by Boyle and Lippman in 1978, and the octree color quantization algorithm [7]. Other quantization methods in the literature are based on histogram analysis [8][9][10], fuzzy logic [11,12], neural network [13,14] and multiresolution analysis [9,10,15].…”

Estimation of Appropriate Parameter Values for Document Binarization Techniques

“…The Adaptive Logical Level Technique (ALLT) and its improvement versions [15,16,29] and the Improvement of Integrated Function Algorithm (IIFA) [17,18,29] are two of the most powerful techniques that use these characteristics. Finally there are binarization techniques that are based on general clustering approaches such as the Fuzzy C-means algorithm (FCM) [19] and the Kohonen neural network-based techniques proposed by Papamarkos et al [20][21][22][23].…”

Color Reduction Using the Combination of the Kohonen Self-Organized Feature Map and the Gustafson-Kessel Fuzzy Algorithm

“…Techniques in this category attempt to find the optimal palette using vector classifiers like the Growing Neural Gas (GNG) [6], Adaptive Color Reduction [7], FOSART [8][9][10][11], Fuzzy ART [12,13] and FCM [14].…”

“…It is preferable to have as training samples data a sub-sampling version of the original image instead of the whole image in order to achieve reduction of the computational time. In the proposed color reduction technique, the training samples are selected in a similar way as in ACR technique [7] from the peaks of the well-known Hilbert's space filling curve [20]. As the Fig.…”