A. C. S. Rao’s research while affiliated with Indian Institute of Technology Dhanbad and other places

Medical imaging is a very important field for diagnosing the diseases from the analysis of x ray, computed tomography (CT) scan or other medical images. Computer aided diagnosis (CAD) helps to physicians to make the clinical decision about the diseases. The most important work in the CAD system to identify the diagnosis details about the images that are used as an input. In this work a computer aided diagnosis system is presented that is used for diagnosing the stages of Lung Cancer by taking Lung ct slices as an input. Pathology relevant regions are called Region of interests (ROI) in Lung ct slices. Region of interests are identified from every ct slices. Features are extracted from the every Region of interests to generate feature vectors. The feature vectors are stored in the database. Based on the Extracted feature vectors training is performed on the Support vector machine classifier (SVM) and after the classification accuracy is evaluated by using the test set.

Segmentation of anatomical structures from chest x ray has an increasing importance in the past four decades and researchers have proposed various techniques and evaluated them using different datasets. In order to evaluate and compare a proposed technique, it is necessary to have knowledge about public datasets available. In this survey, properties and characteristics of different public chest x ray datasets available for segmentation of anatomical structures are studied. Different approaches for segmentation of anatomical structures (lung, heart, clavicles) are summarized. Segmentation techniques for each anatomical structure for a given dataset are compared and analyzed. The paper outlines the issues where further research can be focused.

Segmentation of lungs from chest x ray is a non trivial task required as a preprocessing step for detection of different diseases like cardiomelagy, tuberculosis, pneumonia. High accuracy in segmentation of lung results in high accuracy of detection of diseases from lungs. For the past four decades multiple techniques were proposed for automatic segmentation of lungs. In this paper, we propose a hybrid segmentation technique based on firefly optimized fuzzy c-means clustering algorithm. The output of the fuzzy c-means is given to level set to finalize the segmentation of the lungs. The performance of the proposed technique is evaluated using two public chest x ray datasets: JRST and Montgomery County. JRST contains 247 chest x-rays and MC dataset contains 138 chest x-rays. The Jaccard coefficient for the proposed segmentation technique is 95.1 which is on par with the state of art segmentation techniques.

Landslides have been one of the most damaging natural hazards in the hilly region, which cause loss of life and infrastructure, and hence, landslide susceptibility zonation (LSZ) maps are inevitable for the pre-identification of vulnerable slopes and for the future planning and mitigation programmes. In this study, an integrated remote sensing and geographic information system approach is adopted for the generation of LSZ Map for the Darjeeling and Kalimpong district, West Bengal, India. Topographic maps, satellite data, other informative maps and statistics were utilized. For this study, the causative factors which cause instability of slope such as drainage, lineament, slope, rainfall, earthquake, lithology, land use, geomorphology, soil, aspect and relief were considered. For the generation of LSZ map, thematic data layers were evaluated and generated by assigning appropriate numerical values for each factor weight and their corresponding class rating in the GIS environment. Resulting LSZ map outlines the total study area into five different susceptibility classes: very high, high, moderate, low and very low. This study also demonstrates the classification and prediction of landslide-susceptible zones in coalition with GIS output by using particle swarm optimization–support vector machine approach without feature selection and ant colony optimization approach with feature selection along with support vector machine classifier. GIS-based LSZ map was validated by comparing the landslide frequencies in between the susceptible classes. The usefulness of the LSZ map was also validated by the statistical Chi-square test.

Landslide susceptibility map aids decision makers and planners for the prevention and mitigation of landslide hazard. This study presents a methodology for the generation of landslide susceptibility mapping using remote sensing data and Geographic Information System technique for the part of the Darjeeling district, Eastern Himalaya, in India. Topographic, earthquake, and remote sensing data and published geology, soil, and rainfall maps were collected and processed using Geographic Information System. Landslide influencing factors in the study area are drainage, lineament, slope, rainfall, earthquake, lithology, land use/land cover, fault, valley, soil, relief, and aspect. These factors were evaluated for the generation of thematic data layers. Numerical weight and rating for each factor was assigned using the overlay analysis method for the generation of landslide susceptibility map in the Geographic Information System environment. The resulting landslide susceptibility zonation map demarcated the study area into four different susceptibility classes: very high, high, moderate, and low. Particle Swarm Optimization-Support Vector Machine technique was used for the prediction and classification of landslide susceptibility classes, and Genetic Programming method was used to generate models and to predict landslide susceptibility classes in conjunction with Geographic Information System output, respectively. Genetic Programming and Particle Swarm Optimization-Support Vector Machine have performed well with respect to overall prediction accuracy and validated the landslide susceptibility model generated in the Geographic Information System environment. The efficiency of the landslide susceptibility zonation map was also confirmed by correlating the landslide frequency between different susceptible classes.