A. H. Dekker. Kohonen neural networks for optimal color quantization. Network: Computation in Neural Systems, 5:351--367, 1994.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....by #n# 0 exp # 2 . Again, # 2 is a time constant and # 0 is the initial value of # at the beginning of the SOM algorithm. The distance d j;i between neurons i and j is calculated on the lattice. The SOM network has been employed for bilevel thresholding [16] and for color quantization [17] A fuzzy SOM network thresholds gray level images [18] and a combination of SOM and PCA networks is used for multilevel thresholding in [19] and [20] TASOM (Time Adaptive SOM) networks are modified forms of SOM networks. Each neuron in a TASOM network has its own learning rate and neighborhood ....

....of the image to be segmented. The gray levels are used to train one dimensional weight vectors of the neural network. The authors are with the Computer Engineering Department, Amirkabir University of Technology, Tehran 15914, Iran. E mail: haamed, safa ce.aut.ac.ir. Manuscript received 17 Jan. 2001; revised 14 Aug. 2001; accepted 19 Mar. 2002. Recommended for acceptance by A. Khotanzad. For information on obtaining reprints of this article, please send e mail to: tpami computer.org, and reference IEEECS Log Number 113496. 0162 8828 02 17.00 # 2002 IEEE The proposed GTASOM ....

[Article contains additional citation context not shown here]

A.H. Dekker, "Kohonen Neural Networks for Optimal Color Quantization, " Network: Computation in Neural Systems, vol. 5, pp. 351-367, 1994.

....an image with a limited number of colors. The most usually used techniques for color reduction in a digital image are the color quantization and the multithresholding approaches. The color quantization techniques group similar colors and replace them with only a single quantized color [1] [5], 6] 8] The ultimate goal of the classical color quantization techniques is to reduce the number of colors of an image with minimum distortion [28] Therefore, the main objective of computer graphics research in the color quantization area is to select Manuscript received November 13, 2000; ....

....[23] In this category belong the methods of octree [1] 6] median cut (MC) 8] and variance based algorithm [30] In another major class of color quantization algorithms belong methods based on cluster analysis. The frequently clustering techniques used in this category are the Kohonen SOFM [5], Fuzzy C means [14] 24] C means [24] and K means [29] The above techniques are suitable for eliminating the uncommon colors in an image but they are ineffective for image analysis and segmentation. The quality of the resultant image varies depending on the number of the final colors and the ....

[Article contains additional citation context not shown here]

A. H. Dekker, "Kohonen neural networks for optimal color quantization, " Network: Computat. Neural Syst., vol. 5, pp. 351--367, 1994.

....formation algorithms from regular vector quantization schemes like k means clustering. Topography can be exploited to reduce the consequences of noise in a transmission of quantized data [7, 8, 9, 10] to interpolate between data points [11] to improve the visualization of high dimensional data [12], or in many other applications [3] which make use of distance relations between data points. Even though many people seem to share a common intuitive understanding of what is meant by topography of maps, it has turned out to be difficult to pinpoint this notion in an unequivocal and ....

A. H. Dekker (1994). Kohonen Neural Networks for Optimal Color Quantization. Network 5, 351367.

....formation algorithms from regular vector quantization schemes like k means clustering. Topography can be exploited to reduce the consequences of noise in a transmission of quantized data [7] 8] 9] to interpolate between data points [10] to improve the visualization of high dimensional data [11], or in many other applications [3] which make use of distance relations between data points. Even though many people seem to share a common intuitive understanding of what is meant by topography of maps, it has turned out to be difficult to pinpoint this notion in an unequivocal and ....

A. H. Dekker, Kohonen Neural Networks for Optimal Color Quantization. Network 5, 351-367, 1994.

....region (which is easy to visually monitor) or where the neuron index of the winning node is transformed into a grayvalue or color index in a systematic way. The advantage of neighborhood preservation in false gray value representation of color quantized images was recently highlighted by Dekker [26]. The role of neighborhood preservation in the reconstruction of principal manifolds in the data by interpolation between the best matching node and its neighbors has already been mentioned in the introduction. This argument does not only apply to sophisticated interpolation schemes like the PSOM ....

A. H. Dekker, Kohonen neural networks for optimal color quantization, Network 5, 351-367 (1994).

No context found.

A. H. Dekker. Kohonen neural networks for optimal color quantization. Network: Computation in Neural Systems, 5:351--367, 1994.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC