Electronic Letters on Computer Vision and Image Analysis (ELCVIA - Universitat Autònoma de Barcelona)
Not a member yet
343 research outputs found
Sort by
Finding Kinematic Structure in Time Series Volume Data
Get PDFThis paper presents a new scheme for acquiring 3D kinematic structure and motion from time seriesvolume data. Our basic strategy is to first represent the shape structure of the target in each frame by Reebgraph which we compute by using geodesic distance of target’s surface, and then estimate the kinematicstructure of the target which is consistent with these shape structures. Although the shape structures can bevery different from frame to frame, we propose to derive a unique kinematic structure by way of clusteringsome nodes of graph, based on the fact that they are partly coherent to a certain extent of time series. Oncewe acquire a unique kinematic structure, we fit it to other Reeb graphs in the remaining frames, and describethe motion throughout the entire time series. The only assumption we make is that human body can beapproximated by an articulated body with certain numbers of end-points and branches. We demonstrate theefficacy of the proposed scheme through some experiments.Key Words: Vision-Based Motion Capture, Video and Image Sequence Analysis, Reeb graph, Motion Tracking and Analysis
An Attention Module for Object Detection in Cluttered Images
Get PDFIn this paper, we propose a visual attention module that automatically detects the regions of an input previously unseen image, which are more likely occupied by a known object. The module can be integrated in many object recognition systems for reducing the image space in which to search the object and the computational costs. The proposed strategy has been tested on two public real-world image databases showing good performances and it has been applied to the SIFT recognition algorithm
Human Shape-Motion Analysis In Athletics Videos for Coarse To Fine Action/Activity Recognition Using Transferable Belief Model
Get PDFWe present an automatic human shape-motion analysis method based on a fusion architecture for humanaction and activity recognition in athletic videos. Robust shape and motion features are extracted fromhuman detection and tracking. The features are combined within the Transferable Belief Model (TBM)framework for two levels of recognition. The TBM-based modelling of the fusion process allows to takeinto account imprecision, uncertainty and conflict inherent to the features. First, in a coarse step, actions areroughly recognized. Then, in a fine step, an action sequence recognition method is used to discriminate activities.Belief on actions are made smooth by a Temporal Credal Filter and action sequences, i.e. activities,are recognized using a state machine, called belief scheduler, based on TBM. The belief scheduler is alsoexploited for feedback information extraction in order to improve tracking results. The system is tested onreal videos of athletics meetings to recognize four types of actions (running, jumping, falling and standing)and four types of activities (high jump, pole vault, triple jump and long jump). Results on actions, activitiesand feedback demonstrate the relevance of the proposed features and as well the efficiency of the proposedrecognition approach based on TBM.Key Words: Video Analysis, Human Tracking, Action and Activity Recognition, Transferable Belief Model
Automatic Abdominal Organ Segmentation from CT images
Get PDFIn the recent years a great deal of research work has been devoted to the development of semi-automaticand automatic techniques for the analysis of abdominal CT images. Some of the current interests are theautomatic diagnosis of liver, spleen, and kidney pathologies and the 3D volume rendering of the abdominalorgans. The first and fundamental step in all these studies is the automatic organs segmentation, that is stillan open problem. In this paper we propose our fully automatic system that employs a hierarchical graylevel based framework to segment heart, bones (i.e. ribs and spine), liver and its blood vessels, kidneys, andspleen. The overall system has been evaluated on the data of 100 patients, obtaining a good assessment bothby visual inspection by three experts, and by comparing the computed results to the boundaries manuallytraced by experts
ROI Based Quality Access Control of Compressed Color Image using DWT via Lifting
Get PDFROI (Region of Interest) in an image or video signal contains important information and may be used for access control at various qualities using multiresolution analysis (MRA). This paper proposes quality access control method of compressed color image by modulating the coefficients of ROI at various levels. Data modulation causes visual degradation in the original image and plays the key role in access control through reversible process. The modulation information, in the form of a secret key, is embedded in non-ROI part of the chrominance blue (Cb) channel of the color image using quantization index modulation (QIM). Lifting based DWT, rather than conventional DWT, is used to decompose the original image in order to achieve twofold advantages of (1) low loss in image quality due to QIM and (2) better decoding reliability. Only the authorized user having the full knowledge of the secret key restores the full quality of ROI. Simulation results duly support this claims
Computing von Kries Illuminant Changes by Piecewise Inversion of Cumulative Color Histograms
Get PDFWe present a linear algorithm for the computation of the illuminant change occurring between two colorpictures of a scene. We model the light variations with the von Kries diagonal transform and we estimate itby minimizing a dissimilarity measure between the piecewise inversions of the cumulative color histogramsof the considered images. We also propose a method for illuminant invariant image recognition based onour von Kries transform estimate
Depth Recovery of Complex Surfaces from Texture-less Pair of Stereo Images
Get PDFIn this paper, a novel framework is presented to recover the 3-D shape information of a complex surface using its texture-less pair of stereo images. First a linear and generalized Lambertian model is proposed to obtain SfS data using an image from stereo pair. Then this SfS data is corrected by integrating SIFT indexes. These SIFT indexes are defined by means of disparity between the matching points (scale invariant features) in rectified stereo images. The integration process is based on correcting the 3-D visible surfaces obtained from SfS using these SIFT indexes. The SIFT indexes based improvement of depth values which are obtained from generalized Lambertian reflectance model is performed by a feed-forward neural network. The experiments are performed to demonstrate the usability and accuracy of the proposed framework
Distortion Correction for 3D Scan of Trunk Swaying Human Body Segments
Get PDFWe propose a method for acquiring a 3D shape of human body segments accurately. Using a light stripe
triangulation range finder, we can acquire accurate 3D shape of a motionless object in dozens of seconds.
If the object moves during the scanning, the acquired shape would be distorted. Naturally, humans move
slightly for making balance while standing even if the subject makes an effort to stay still for avoiding the
distortion in acquired shape. Our method corrects the distortion based on measured subject’s motion during
the scanning.
Experimental results show the accuracy of the proposed method. Trunk swaying degrades the accuracy
of the light stripe triangulation from 1mm to 10mm. We can keep the accuracy of as good as 2mm by
applying our method
Complex networks : application for texture characterization and classification
Get PDFThis article describes a new method and approch of texture characterization. Using complex network representation of an image, classical and derived (hierarchical) measurements, we presente how to have good performance in texture classification. Image is represented by a complex networks : one pixel as a node. Node degree and clustering coefficient, using with traditionnal and extended hierarchical measurements, are used to characterize ”organisation” of textures
Neighbouring Color Dependence Matrix for Image Analysis : Application to homogeneous and heterogeneous areas detection and characterization
Get PDFA new method for color texture characterization and color texture region detection is presented. This method, which we will name NCDM (Neighbouring Color Dependence Matrices), is the extension to color textures of the NGLDM (Neighbouring Gray Level Dependence Matrices) introduced by Sun et al. [1] and completed by Berry et al. [2]. This approach consists in estimating the dependences of colors between a pixel and its neighbours. We propose two steps: a color areas classification in two classes followed by the characterization of the detected areas. In the first step, we compute the NCDM with an isotropic neighbourhood. The structure of the isotropic NCD distribution allow us to separate the pixels of a color composite image into two classes, which correspond respectively to homogeneous and heterogeneous regions in the image. We then consider that the heterogeneous regions are potentially textured regions and in the second step we propose to compute the NCDM with anisotropic neighbourhoods corresponding to the eight principal directions. To seek the dominant directions in a color texture, a measure of spatial dependence between a pixel and its neighbours is computed by way of a chi-square test. This measure is based on the fit of the NGLD and NCD distribution with a binomial model under independence hypothesis. The variations of the colors are computed in uniform perceptual color spaces. We have chosen the color space ";L1 norm"; introduced by Angulo and Serrakeywords: color space, anisotropy, NGLD