Fig 1 - uploaded by Songul Varlı Albayrak
Content may be subject to copyright.
The quantisation results for 1-D (a) and 2-D Self-Organizing Feature Map algorithm (b) of image Peppers(c) with 3000 iteration.
Source publication
Color quantization process is performed by clustering in color space. The clustering algorithm we examine is self-organizing feature map (SOFM) introduced by Kohonen. In this application we use a one- and two-dimensional self-organizing neural network and compare them. In the competitive learning process, the weigh vectors for each neuron are produ...
Similar publications
EXTENDED ABSTRACT In this paper, we use Self-Organising Maps (Kohonen 1989) together with a helical encoding of time. The maps can therefore learn the periodic variation in input data, which may be annual, weekly, or daily. We have implemented this technique in software, and have tested it by demonstrating how it can be used to predict missing data...
Many data sets contain an abundance of background data or samples belonging to classes not currently under consideration. We present a new unsupervised learning method based on Fuzzy C-Means to learn sub models of a class using background samples to guide cluster split and merge operations. The proposed method demonstrates how background samples ca...
Citations
... These nodes are arranged in either a 1D or a 2D lattice. The choice of 2D lattice is recommended by the SOM methodology because 1D lattices are insufficient in capturing features of high dimensional input spaces 49 (Albayrak, 2002). SOM creates topologically ordered mappings between input data, and the nodes of the map. ...
... Rather, it is just a lattice on which neurons representative of parts of composition space reside. The choice of a 2D lattice is recommended by the SOM methodology-the employment of an 1D lattice is not sufficient to capture the feature of high dimensional input spaces, as was shown by Albayarak [30] , and no need for employing higherdimensional lattices has been identified. ...
In this work, a methodology for the tabulation of combustion mechanisms via Artificial Neural Networks (ANNs) is presented. The objective of the methodology is to train the ANN using samples generated via an abstract problem, such that they span the composition space of a family of combustion problems. The abstract problem in this case is an ensemble of laminar flamelets with an artificial pilot in mixture fraction space to emulate ignition, of varying strain rate up to well into the extinction range. The composition space thus covered anticipates the regions visited in a typical simulation of a non-premixed flame. The ANN training consists of two-stage process: clustering of the composition space into subdomains using the Self-Organising Map (SOM) and regression within each subdomain via the multilayer Perceptron (MLP). The approach is then employed to tabulate a mechanism of CH4–air combustion, based on GRI 1.2 and reduced via Rate-Controlled Constrained Equilibrium (RCCE) and Computational Singular Perturbation (CSP). The mechanism is then applied to simulate the Sydney flame L, a turbulent non-premixed flame that features significant levels of local extinction and re-ignition. The flow field is resolved through Large Eddy Simulation (LES), while the transported probability density function (PDF) approach is employed for modelling the turbulence–chemistry interaction and solved numerically via the stochastic fields method. Results demonstrate reasonable agreement with experiments, indicating that the SOM-MLP approach provides a good representation of the composition space, while the great savings in CPU time allow for a simulation to be performed with a comprehensive combustion model, such as the LES-PDF, with modest CPU resources such as a workstation.
... In post-clustering approach, small numbers of Cluster centres are selected randomly and each colour is placed in a Cluster corresponding to which they are closest. The typical clustering algorithms, such as K-means clustering algorithm [13] and Self-Organising Feature Map (SOFM) [14], can be grouped as post-clustering approaches. By training samples, the quantized colours that are sometimes called codebook or lookup table can represent a colour image better. ...
