Browse Journal: IST Transactions of Computer Vision Systems-Theory and Applications

As fingerprint verification depends strongly on the quality of fingerprint ridge orientation estimation, so the more accurate fingerprint ridge orientations estimate, the better verification will result in. In this paper, we have proposed a new technique for improving the coarse fingerprint orientation estimation smoothing using fingerprint orientation model based on 2D Fourier expansion with a special Modification on it (M-FOMFE). The modification we have used in this paper is taking into account the information of Coherence Matrix for fingerprint ridge orientation estimation. This matrix is used to retrieve the uncertainty of each ridge orientation block and improves the accuracy of old FOMFE coarse ridge orientation smoothing method. Moreover, a Gaussian filter has been combined with the proposed M-FOMFE to improve the ridge orientation field at local regions. For evaluating the proposed methods, reliable and robust evaluating methods have been employed such as a fingerprint continuous classification, an exclusive one based on Poincare-Index algorithm for singular point detection and also a performance measure based on the rate of SP position detection accuracy. Compared to competing methods, the experimental results show that the proposed methods have better orientation estimation, classification and SP detection performances.

Cardiac Magnetic Resonance Imaging (MRI) is a non invasive medical test that helps physicians diagnose and treat medical conditions. A single cardiac examination results in a large amount of image data. Manual analysis by experts is time consuming and also susceptible to intra-observer and inter-observer variability. This leads to the urgent requirement for efficient image segmentation algorithms to automatically extract clinically relevant parameters. We propose an active contour model to have accurate left ventricle (LV) segmentations across a cardiac cycle. This paper will facilitate the segmentation and analysis of dynamic images by saving them as frames. The segmentation at each particular time is based not only on the data observed at that instant, but also by using the boundaries tracked in the previous frame as initial frame. The framework uses a curve evolution method along with energy inimization to estimate the LV boundaries at each time. Further, the volumetric analysis of the segmented image will give us the quantity of blood that enters the left ventricle in one cardiac cycle and the clinical parameters such as end systolic volume, end diastolic volume and ejection fraction can be calculated.

We describe a texture based method for classification of exudates from Diabetic retinopathy (DR) images. This is realized by extracting twelve features from 1305 samples of both normal and pathological retinal images. The proposed method is tested with kNN classifier and with feedforward neural network and three variations of backpropogation algorithm.