To search, Click below search items.


All Published Papers Search Service


An Identification of Abnormal Images using semi-supervised Segmentation Model


V.Manogajapathi, M.Frankline, R.Karthikeyen


Vol. 15  No. 2  pp. 98-102


The objective of semi-supervised image segmentation is to obtain the segmentation from a partially labeled image. By exploit the image numerous structure in labeled and unlabeled pixels, semi-supervised methods propagates the user labelling to unlabeled data, thus reducing the need for the user labeling. Several semi-supervised research mechanism have been proposed in the observation. In this paper, we consider the offending of segmentation of large collections of images and the classification of images by related diseases. We are detecting abnormal images by the process of segmentation and classification. The segmentation used in this paper has two advantages. First, user can specify their own values by highly controlling the segmentation. Another is, at initial stage this mechanism needs only minimum tuning of model parameters. Once initial tuning process is done, the setup can be used to automatic segment a large collection of images that are distinct but share identical features. And for classification of diseases, a numerous research method, called parameter-free semi-supervised local Fisher discriminant analysis is used. This method preserves the global structure of labelled samples in addition to separating unlabelled samples in various classes from each other. The semi-supervised method has a systematic form of the globally optimal solution, which can be computed efficiently by Eigen decomposition. Espousal experiments on various collections of biomedical images suggest that the proposed mechanism is effective for segmentation with classification and is computationally adequate


Biological image segmentation, semi-supervised segmentation, multiple imaging, ,microscopy images