Lung cancer accounts for the highest number of cancer deaths globally. Early diagnosis of lung nodules is very important to reduce the mortality rate of patients by improving the diagnosis and treatment of lung cancer. This work proposes an automated system to classify lung nodules as malignant and benign in CT images. It presents extensive experimental results using a combination of geometric and histogram lung nodule image features and different linear and non-linear discriminant classifiers. The proposed approach is experimentally validated on the LIDC-IDRI public lung cancer screening thoracic computed tomography (CT) dataset containing nodule level diagnostic data. The obtained results are very encouraging correctly classifying 82% of malignant and 93% of benign nodules on unseen test data at best.
Broadly speaking, the objective in cardiac image segmentation is to delineate the outer and inner walls of the heart to segment out either the entire or parts of the organ boundaries. This paper will focus on MR images as they are the most widely used in cardiac segmentation -- as a result of the accurate morphological information and better soft tissue contrast they provide. This cardiac segmentation information is very useful as it eases physical measurements that provides useful metrics for cardiac diagnosis such as infracted volumes, ventricular volumes, ejection fraction, myocardial mass, cardiac movement, and the like. But, this task is difficult due to the intensity and texture similarities amongst the different cardiac and background structures on top of some noisy artifacts present in MR images. Thus far, various researchers have proposed different techniques to solve some of the pressing issues. This seminar paper presents an overview of representative medical image segmentation techniques. The paper also highlights preferred approaches for segmentation of the four cardiac chambers: the left ventricle (LV), right ventricle (RV), left atrium (LA) and right atrium (RA), on short axis image planes.
This work presents an automatic human gender and age group recognition system based on human facial images. It makes an extensive experiment with row pixel intensity valued features and Discrete Cosine Transform (DCT) coefficient features with Principal Component Analysis and k-Nearest Neighbor classification to identify the best recognition approach. The final results show approaches using DCT coefficient outperform their counter parts resulting in a 99% correct gender recognition rate and 68% correct age group recognition rate (considering four distinct age groups) in unseen test images. Detailed experimental settings and obtained results are clearly presented and explained in this report.