Performing the segmentation of vasculature in the retinal images having pathology is a challenging problem. This paper presents a novel approach for automated segmentation of the vasculature in retinal images. The approach uses the intensity information from red and green channels of the same retinal image to correct non-uniform illumination in color fundus images. Matched filtering is utilized to enhance the contrast of blood vessels against the background. The enhanced blood vessels are then segmented by employing spatially weighted fuzzy c-means clustering based thresholding which can well maintain the spatial structure of the vascular tree segments. The proposed method's performance is evaluated on publicly available DRIVE and STARE databases of manually labeled images. On the DRIVE and STARE databases, it achieves an area under the receiver operating characteristic curve of 0.9518 and 0.9602 respectively, being superior to those presented by state-of-the-art unsupervised approaches and comparable to those obtained with the supervised methods.

An efficient approach for automatic detection of red lesions in ocular fundus images based on pixel classification and mathematical morphology is proposed. Experimental evaluation of the proposed approach demonstrates better performance over other red lesion detection algorithms, and when determining whether an image contains red lesions the proposed approach achieves a sensitivity of 100% and specificity of 91%.

No abstract

The accurate land use land cover (LULC) classifications from satellite imagery are prominent for land use planning, climatic change detection and eco-environment monitoring. This paper investigates the accuracy and reliability of Support Vector Machine (SVM) classifier for classifying multispectral image of Hyderabad and its surroundings area and also compare its performance with Artificial Neural Network (ANN) classifier. In this paper, a hybrid technique which we refer to as Fuzzy Incorporated Hierarchical clustering has been proposed for clustering the multispectral satellite images into LULC sectors. INTRODUCTIONIn recent times, satellite images are delivering an enormous source of information for studying the spatial and temporal variability of environmental conditions. It can be utilized in a number of applications which consists of making of mapping products for military and civil purposes, exploration, nursing of land use land cover [1], assessment of environmental damage, radiation level check, soil test, growth directive, crop outcome increment and urban planning. These multispectral images can be used mainly in the course of classification and also in mapping of vegetation over extensive spatial scales. This is because multispectral image classifies land cover and land usage features such as vegetation, water, oil, forests and urbanization, it also delivers very good scope and mapping. This kind of classification technique replaces the traditional classification techniques [2]. As many environmental and socio-economic proposals are based on these classification results, researches and studies on these satellite image classifications have centralized the concentration of the scientific community. A classification technique classifies relevant components or classes of land cover sections over a particular area [3]. Multi spectral images primarily consists the data that is collected over a wide range of frequencies and it changes for different locations [4]. This overall peculiar nature of satellite image data can be related to the spectral features that correlate with the spatial features belonging to the same band and this is called as spatial correlation. Classification of land use and land cover using remote sensing imagery is tough because of compound landscapes and also the spectral and spatial resolution of the satellite images.

No abstract