Image Classification Method Based on Rough Set
|Course||Operational Research and Cybernetics|
|Keywords||Image classification Rough sets Variable precision rough set model Bayesian rough set model|
Image classification information is an important research direction. In this paper, the image classification involved in the key technologies, including image feature extraction, image data to establish the decision-making table, based on rough set theory attributes the reduction algorithm, rules reduction algorithm, sample prediction algorithm, etc. which have been studied and discussed, through the final adoption test and evaluation, is access to the accurate classification accuracy.Rough set model by the mathematician Pawlak first is proposed to deal with a vague and uncertain knowledge of the new mathematical tools, following with the rough set the continuous in-depth research, the limitations of rough set model has been gradually revealed, the use of rough handling of classified set be completely correct or yes. which caused a lot of useful information in the process of extracting the rules lost, the rough set of limitations restricting its application. Thus, in recent years, many scholars from various aspects of rough set model to promote, a variable precision rough set model, the probability of rough set model, generalized rough set model, fuzzy rough set model.Variable precision rough set model is the standard rough set model introduced in the correct classification rate, easing the standard rough set model similar to the rigorous demands of the border; strengthen the rough set model of anti-jamming capabilities and new data on the predictive capability.This article is based on variable precision rough set model and Bayesian rough set model; the image classification problems were studied. Firstly, the title introduces rough set mainly on theoretical research profile and image classification issue of the status quo, shows rough set the standard model, information systems, attribute reduction and variable precision rough set model. Bayesian rough set model of the basic concepts. Secondly, the title mainly introduces the image classification problem based on color characteristics, giving the common color space conversion and mutual methods, mainly in the classic rough set on the basis of classification model using variable precision rough set model, the introduction of similar distinction. It makes the concept of Matrix, a color based on the characteristics of the image classification model and its classification algorithm, through the example of classification of the effectiveness and feasibility. Thirdly, through the introduction of the overall gain and relatively Bayesian distinction matrix, Bayesian rough set model, which is given two attribute reduction algorithm.lastly, the relative gain from an overall perspective on the important attributes of the analysis and information is given as a heuristic Bayesian rough set attribute reduction of heuristics,and, the distinction between the introduction of Bayesian matrix, the frequency of use of properties and properties as the length of stimulating factors, and this is a rough set Bayesian attribute reduction of another algorithm, the final presents a color feature Bayesian image classification model and its classification algorithm. The results show that Bayesian rough set has a practical value in the image classification issue, the Bayesian approach rough set of good, accurate and efficient classification.