Publications

Abstract : Synchronous data processing is often preferred considering the volume of data available and the requirement of processed data. Data volume increase from current technical advancements should be balanced and suitably dealt with efficient feature processing techniques. Feature selection can efficiently improve the concurrency in data processing on a completely uncorrelated and separately processable feature set. Texture analysis is a significant application of image processing that analyses specific patterns in an image. As the volume of image data available for producing better classification models progressively moves to be out of the human-manageable range, it is the automation methods that aid the researchers to manage the data. When the number of samples in the data and the dimensionality of the problem is too high, the purpose of data analysis is often compromised. Feature engineering can be utilized to improve the efficiency of the model with a smaller subset of features. This work proposes dibyhrid bio-inspired computation based feature compaction, that is a combination of improvised genetic algorithm and selective tabu search for texture classification problem. Machine learning techniques are employed to evaluate and minimize the validation error from the applied computational methods.