Improving Feature Selection Capabilities in Skin Disease Detection System
Vedanti Chintawar1, Jignyasa Sanghavi2
1Vedanti Chintawar, M.Tech Scholar, Nagpur University, Ramdeobaba college of Engineering & Management, Nagpur, India.
2Jignyasa Sanghavi, Assistant professor, Nagpur University, Ramdeobaba college of Engineering & Management, Nagpur, India.
Manuscript received on 08 June 2019 | Revised Manuscript received on 13 June 2019 | Manuscript Published on 08 July 2019 | PP: 247-251 | Volume-8 Issue-8S3 June 2019 | Retrieval Number: H10680688S319/19©BEIESP
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open-access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Feature extraction is the process of description of the input imagery into a fixed set of values. For a good feature extraction algorithm these values are sufficient in order to describe the entire properties of the image under test. There are many kind of features which can be extracted from the image, these features vary from color features, shape features, to morphological and texture features. Feature extraction is usually application dependent, and allows the application designers to incorporate various kinds of descriptors for the image under test. An optimum feature extraction system is the one which can accurately and uniquely identify each image separately via uniqueness in the feature vector. In this paper, we analyze various feature extraction techniques and identify the best features suited for the application of skin disease detection systems, and also provide some acute observations on how these techniques can be improved to further optimize the accuracy of identification.
Keywords: Accuracy, color, feature, morphological, shape, texture, uniquely.
Scope of the Article: Artificial Intelligence