Niloofar Gheissari

Segmentation of Dermoscopy Images Using Wavelet Networks

This paper introduces a new approach for the segmentation of skin lesions in dermoscopic images based on wavelet network (WN). The WN presented here is a member of fixed-grid WNs that is formed with no need of training. In this WN, after formation of wavelet lattice, determining shift and scale parameters of wavelets with two screening stage and selecting effective wavelets, orthogonal least squares algorithm is used to calculate the network weights and to optimize the network structure. The existence of two stages of screening increases globality of the wavelet lattice and provides a better estimation of the function especially for larger scales. R, G, and B values of a dermoscopy image are considered as the network inputs and the network structure formation. Then, the image is segmented and the skin lesions exact boundary is determined accordingly. The segmentation algorithm were applied to 30 dermoscopic images and evaluated with 11 different metrics, using the segmentation result obtained by a skilled pathologist as the ground truth. Experimental results show that our method acts more effectively in comparison with some modern techniques that have been successfully used in many medical imaging problems.

Estimating the Mutual Information Between Bilateral Breast in Thermograms Using Nonparametric Windows

Comparison between contra lateral breast images is one of the effective methods in breast cancer detection. Asymmetric temperature distribution can be an indicator of abnormality. The mutual information is a good measure of nonlinear correlation. It is a measure that captures linear and nonlinear dependencies, without requiring the specification of any kind of model of dependence. Therefore, it is suitable for our abnormality indicator. Although nonparametric windows is a numerically expensive technique but it is accurate. The reason is that nonparametric windows incorporate an interpolation model which enhances the resolution to a highly oversampled image. For our purposes we worked with sixty simulated breast thermal images. It is shown that the more similar the thermal image of right breast to the thermal image of left breast, the closer the normalized mutual information value to one.

A new invariant descriptor for action recognition based on spherical harmonics

The aim of this paper is to introduce a new descriptor for the spatio-temporal volume (STV). Human motion is completely represented by STV (action volume) which is constructed over successive frames by stacking human silhouettes in consecutive frames. Action volume comprehensively contains spatial and temporal information about an action. The main contribution of this paper is to propose a new affine invariant action volume descriptor based on a function of spherical harmonic coefficients. This means, it is invariant under rotation, non-uniform scaling and translation. In the 3D shape analysis literature, there have been a few attempts to use coefficients of spherical harmonics to describe a 3D shape. However, those descriptors are not affine invariant and they are only rotation invariant. In addition, the proposed approach employs a parametric form of spherical harmonics that handles genus zero surfaces regardless of whether they are stellar or not. Another contribution of this paper is the way that action volume is constructed. We applied the proposed descriptor to the KTH, Weizmann, IXMAS and Robust datasets and compared the performance of our algorithm to competing methods available in the literature. The results of our experiments show that our method has a comparable performance to the most successful and recent existing algorithms.

Human motion analysis using dynamic textures

Human motion analysis is a dynamic field of research in computer vision. Its popularity is because of wide application in surveillance, study of social interaction, robot guidance and video indexing. Human motion can be considered as dynamic texture since it has statistical variation in spatiotemporal domain. The local interest features contain efficient information of these spatiotemporal variations. The proposed method is based on dynamic texture description for analysis of human motion using visual dictionary. This method is applied to two major applications: action recognition and gait recognition. We evaluate the performance of our method on KTH dataset for both action and gait recognition.

Gait recognition using dynamic texture descriptors

The human motion analysis is an attractive topic in biometric research. Common biometrics is usually time-consuming, limited and collaborative. These drawbacks pose major challenges to recognition process. Recent researches indicate people have considerable ability to recognize others by their natural walking. Therefore, gait recognition has obtained great tendency in biometric systems. Gait analysis is inconspicuous, needs no contact, cannot be hidden and is evaluated at distance. This paper presents a bag of word method for gait recognition based on dynamic textures. Dynamic textures combine appearance and motion information. Since human walking has statistical variations in both spatial and temporal space, it can be described with dynamic texture features. To obtain these features, we extract spatiotemporal interest points and describe them by a dynamic texture descriptor. To get more suitable results, we extend LBP-TOP as a rotation invariant dynamic texture descriptor. Afterwards, hierarchical K-means algorithm is employed to map features into visual words. At result, human walking represent as a histogram of video-words occurrences. We evaluate the performance of our method on two dataset: the KTH dataset and IXMAS multiview dataset.

Automatic aerial image segmentation based on a modified Chan-Vese algorithm

Automatic segmentation of aerial images has been a challenging task in recent years. Region-based active contour of Chan-Vese has been proposed to detect objects in a given image. This algorithm is more powerful than classical edge-based active contour algorithms. In this paper, aerial images are automatically segmented into a number of homogeneous areas using Chan-Vese model implemented by Narrow Band Level Set method with reinitialization together with extracting color and texture features. For this purpose, a variety of different color and texture features have been tested. The results show that incorporation of Gabor filters in HSV color space leads the most accurate results.