G. Kavitha

Analysis of vasculature in human retinal images using particle swarm optimization based tsallis multi-level thresholding and similarity measures

Retinal vasculature of the human circulatory system which can be visualized directly provides a number of systemic conditions and can be diagnosed by the detection of lesions. Changes in these structures are found to be correlated with pathological conditions and provide information on severity or state of various diseases. In this work, particle swarm optimization algorithm based multilevel thresholding is adopted for detecting the vasculature structures in retinal fundus images. Initially, adaptive histogram equalization is used for pre-processing of the original images. Tsallis multilevel thresholding is used for the segmentation of the blood vessels. Further, similarity measures are used to quantify the similarity between the segmented result and the corresponding ground truth. The optimal multi-threshold selection using particle swarm optimization seems to provide better results. Similarity measures analysis using dendrogram and box plot provide validation of the segmentation procedure attempted.

An Approach to Identify Optic Disc in Human Retinal Images Using Ant Colony Optimization Method

In this work, an attempt has been made to identify optic disc in retinal images using digital image processing and optimization based edge detection algorithm. The edge detection was carried out using Ant Colony Optimization (ACO) technique with and without pre-processing and was correlated with morphological operations based method. The performance of the pre-processed ACO algorithm was analysed based on visual quality, computation time and its ability to preserve useful edges. The results demonstrate that the ACO method with pre-processing provides high visual quality output with better optic disc identification. Computation time taken for the process was also found to be less. This method preserves nearly 50% more edge pixel distribution when compared to morphological operations based method. In addition to improve optic disc identification, the proposed algorithm also distinctly differentiates between blood vessels and macula in the image. These studies appear to be clinically relevant because automated analyses of retinal images are important for ophthalmological interventions.

Identification and analysis of macula in retinal images using Ant Colony Optimization based hybrid method

In this work, a hybrid approach to analyse optic disc and macula to characterise the normal and abnormal status of the retina are proposed. The retinas with normal and Diabetic Retinopathy (DR) images were used for this study. The fundus retinal images are subjected to Ant Colony Optimization (ACO) based method to identify optic disc (OD) and Otsu method to further analyse the macula. Parameters such as radius of optic disc and distance between the centres of the optic disc and macula are used as indices for evaluation. The results show that combining Otsu method with ACO based method for macula detection demonstrate improved performance than ACO method alone. The value of the radius of optic disc and distance between the centres of OD and macula are distinct for normal and abnormal images. As identification of OD and macula are important for pathological assessment these studies seems to be clinically relevant.

Analysis of ventricle regions in Alzheimer’s brain MR images using level set based methods

In this work, an attempt has been made to analyse ventricle region of the T1 weighted coronal Magnetic Resonance (MR) brain images and study the progression of severity in Alzheimer’s Disease (AD) conditions. Two level set methods namely Distance Regularised Level Set Evolution (DRLSE) and geodesic active contour are used to extract the desired region of interest. Eighty geometric features are derived from the segmented ventricle region. The most significant parameters are found using principal component analysis. Results demonstrate that the DRLSE shows better performance in extraction of the boundary of the ventricle region than geodesic active contour method. The geometrical feature, area is found to have a high correlation with brain to ventricle index for all subjects. Further, it is observed that this feature gives a distinct separation between normal and abnormal AD subjects (p value = 0.00012). It also provides high correlation for normal (.97) and abnormal AD subjects (<0.9). Hence, this analysis could be a useful supplement to physicians in diagnosis and treatment of Alzheimer’s and other neurodegenerative disorders.

Proposal of a Content Based retinal Image Retrieval system using Kirsch template based edge detection

In this work, a Content Based Image Retrieval (CBIR) frame work is developed based on edge detection method for diagnosis of diabetic retinopathy. Normal and abnormal retinal fundus images are subjected to preprocessing methods to enhance the edge information. Two different methods namely Kirsch template and Canny edge based detection techniques are considered for segmentation of blood vessels. The structure and texture based features obtained from segmented images are analyzed. Best features for retinal image retrieval are selected from the quantitative analysis of features. Similarity matching is carried out using Euclidean distance method and the retrieved images are ranked. Retrieval efficiency is calculated in terms of precision and recall. The results show that the Kirsch template based edge detection method identifies most of the blood vessels compared to the other method. High degree of precision and recall are observed using the Kirsch template based CBIR system. It appears that the Kirsch edge based detection could be useful in CBIR system for diagnosis of retinal abnormalities.

Segmentation and grading of diabetic retinopathic exudates using error-boost feature selection method

This paper proposes a method to segment the exudates and lesions in retinal fundus images and classify using selective brightness feature. The exudates are segmented from background and their size is also measured. The segmentation is done by extraction of pixels which fall in the color range of the spots. The essential features inferred from the segmented image include the count of the exudates, maximum size, percentage affected, color intensity of the spot, average size and the area affected by haemorrhages. The diagnosis is supported by error-boost feature selection technique. This technique classifies the retinal images as normal or abnormal based on the features obtained from the segmented image. The abnormal images are further classified as mild, moderate or severe and there is an additional classification based on non-proliferative and severe proliferative diabetic retinopathy. The diagnosis parameter ranges for each feature are set prior to the severity classification. The error boost feature selection algorithm selects the key features which classifies the retinopathy more accurately. The obtained results seem to be clinically relevant.

Detection of blood vessels in human retinal images using Ant Colony Optimisation method

In this, an attempt has been made to analyse blood vessels in human retinal digital images using Ant Colony Optimisation (ACO) based edge detection algorithm and was then correlated with Otsu and Matched filter methods. Results show that the ACO method provides high visual quality output with better detection of blood vessels. It provides better delineation, distinctly differentiates central veins, extracts small blood vessels and detects abnormalities in the image. The ratio of vessel-to-vessel free area using ACO method is distinctly different for normal and abnormal images ( p < 0.005). It appears that this study is useful for mass screening and avoids complications at later stages of diseases.

Analysis of the liver in CT images using an improved region growing technique

This paper presents an improved region growing algorithm to enhance the segmentation of the liver from abdominal CT images. The abdominal CT images are characterized by poor contrast and blurred edges which increase the complexity of liver segmentation. Initially, the images are subjected to preprocessing which involves de-noising, thresholding and non-linear mapping. Then, the improved region growing algorithm is applied to the preprocessed liver images. Post processing is performed using a combination of morphological operations. The results of the improved algorithm are compared with the traditional region growing algorithm and the k-means clustering algorithm to show the effectiveness of the proposed method. Performance validation is also done by comparing the results with the ground truth. Similarity measures namely the Dice similarity, Sokal and Sneath-I similarity, Sokal and Sneath-II similarity and Tanimoto similarity are used for the comparison. The results obtained using the improved method give an accuracy of 97%. The average Dice similarity measure for the considered images was found to be 0.86. The average correlation coefficient between the ground truth and the segmented result are also high in the improved algorithm. The obtained results seem to be clinically relevant.

Analysis of sub-cortical regions in cognitive processing using fuzzy c-means clustering and geometrical measure in autistic MR images

Magnetic Resonance (MR) imaging is an indispensable approach for obtaining the structural information of sub-cortical regions of brain. In this work, fuzzy c-means clustering (FCM) is used to segment the sub-cortical regions of brain such as Corpus Callosum (CC) and Brain Stem (BS). The geometrical measure area is calculated from the extracted regions and correlated with the clinical Intelligent Quotient (IQ) values. The corpus callosum area gives distinct variation between control and autistic subjects (p=0.006). Also, the CC area of autistic subjects is correlated with the verbal IQ value (R=-0.37). The area of brain stem is less statistically significant (p=0.02) compared to the CC area in discriminating the subjects. Also, BS area of autistic subjects gives a correlation of R=-0.29 with the performance IQ. As the reduced CC and BS area are related with cognitive dysfunctions, this framework can be used for the automated diagnosis of autism like neural disorders.

Analysis of Vasculature Detection in Human Retinal Images UsingBacterial Foraging Optimization Based Multi Thresholding

Analysis of blood vessels in digital retinal fundus images is an important problem attempted in contemporary biomedical engineering research. In this work, normal and abnormal retinal images are pre-processed with adaptive histogram equalization and fuzzy filtering. Pre-processed images are then subjected to Tsallis multi-level thresholding method. The threshold levels determined by the chosen method are further optimized using bacterial foraging optimization techniques in order to improve the vessel content. The obtained results are validated using similarity measures by comparing with the corresponding ground truth of each image. Statistical and Tamura features are derived from optimal multi-level thresholding output images to analyse the healthy and pathological images. Results demonstrate that attempted series of pre-processing techniques enhances the edge information considerably and improves the efficacy of segmentation. It is observed that bacterial foraging optimization for Tsallis multi-level thresholding is able to extract retinal vasculature. Similarity measures show that this method provides considerable improvement in the extraction of vessel edges. Further, the statistical and Tamura features derived from detected vessels provide better differentiation between healthy and pathological images. As presence and absence of vessels in retina are clinically significant, the findings seem to be useful.