International Journal of Image and Graphics

Similarity metric is a key component in query-by-example image searching with visual features. After extraction of image visual features, the scheme of computing their similarities can affect the system performance dramatically — the image searching results are normally displayed in decreasing order of similarity (alternatively, increasing order of distance) on the graphical interface for end users. Unfortunately, conventional similarity metrics, in image searching with visual features, usually encounter several difficulties, namely, lighting, background, and viewpoint problems. From the signal processing point of view, this paper introduces a novel similarity metric and therefore reduces the above three problems to some extent. The effectiveness of this newly developed similarity metric is demonstrated by a set of experiments upon a small image ground truth.

We present an automated and complete camera-based monitoring system that makes use of low-level color features to perform detection, tracking and recognition of multiple people in video sequence. Specifically, the system employs a novel coverage check-up method to segment detected foreground regions into isolated people and then localize each of them. During tracking, the appearances of people are modeled by their color histograms so that the system can keep aware of their identities and recognize them after occlusions by maximizing the joint likelihood. To make the recognition more robust against shadows or changes of background illumination, the system also incorporates a shadow removal scheme to suppress shadow effects and hence improve the quality of color histogram. The proposed system has been used to identify people who re-enter the field of view of a monitoring camera in a closed-environment. Experimental results of real video data demonstrate the efficacy of the proposed people monitoring system.

Whole human body scanners are 3D imaging devices which are capable of capturing a computerized format of whole body shape, thus permitting automatic extraction of the different body measurements. This requires the segmentation of scan data into subsets corresponding to the functional human body parts. Such a task is quite challenging due to the articulated and the deformable aspects of the human body shape. The attempts made so far suffer from various limitations, such as being restricted to standard specific posture and vulnerability to scan data corruption. This paper proposes a general framework that aims towards overcoming these challenges. One of the salient features of this framework is that it can cope with moderate posture variations around the standard posture, in addition of being quite robust against noise, holes and irregular sampling. Experimental results performed on real and synthetic data confirmed the validity, effectiveness and robustness of our framework.

The recent increase in the generated polygonal dataset sizes has outpaced the performance of graphics hardware. Several solutions such as multiresolution hierarchies and level-of-detail rendering have been developed to bridge the increasing gap. However, the discrete levels of detail generate annoying popping effects, the preliminaries multiresolution schemes cannot perform drastic changes on the selected level of detail within the span of small number of frames, and the current cluster-based hierarchies suffer from the high-detailed representation of the boundaries between clusters. In this paper, we are presenting a novel approach for multiresolution hierarchy that supports dual paths for run-time adaptive simplification — fine and coarse. The proposed multiresolution hierarchy is based on the fan-merge operator and its reverse operator fan-split. The coarse simplification path is achieved by directly applying fan-merge/split, while the fine simplification route is performed by executing edge-collapse/vertex-split one at a time. The sequence of the edge-collapses/vertex-splits is encoded implicitly by the order of the children participating in the fan-merge/split operator. We shall refer to this multiresolution hierarchy as fan-hierarchy. Fan-hierarchy provides a compact data structure for multiresolution hierarchy, since it stores 7/6 pointers, on the average, instead of 3 pointers for each node. In addition, the resulting depth of the fan-hierarchy is usually smaller than the depth of hierarchies generated by edge-collapse based multiresolution schemes. It is also important to note that fan-hierarchy inherently utilizes fan representation for further acceleration of the rendering process.

In this paper, a new fast one-step codebook design algorithm for vector quantization is proposed for image coding. The algorithm utilizes a fuzzy clustering methodology for best clustering of data vectors in the training space. The codebook design process terminates in just one step so that it is highly computationally efficient as compared to other reported algorithms. To improve the coding efficiency, image blocks are further classified into two classes and a different codebook is designed for each class. The two codebooks are augmented to form one codebook so that there is no need to send class information. Moreover, entropy coding is used to send codevector index to the receiver for further bit rate reduction. Simulation results are presented to show the superior performance of the proposed algorithm in terms of PSNR as compared to the state-of-the-art codebook design algorithms.

A merging procedure joining search pattern and variable block size motion estimation for H.264/AVC is proposed in this paper. The principal purpose of the proposed methods is the reduction of the computational complexity for block matching module. In fact, there are numerous contributions in the literature aiming the reduction of the computational cost needed for motion estimation. The best solution from a qualitative point of view is the full search that considers every possible detail. The computational effort required is enormous and this makes motion estimation by far the most important computational bottleneck in video coding systems. Our approach invests and exploits the center-biased characteristics of the real world video sequences, aiming to achieve an acceptable image quality while independently targeting the reduction of the computational complexity. The simulations results demonstrated that the proposal performs well.

In industry applications, the range images are generally huge points arrays and are additively noised. They usually represent surfaces of 3D objects and are used for reverse engineering process in CAD/CAM domains. To compute the geometrical model of each surface present in the range image, we denoise and sub-sample the raw range data. Denoising allows us to avoid the adverse effects of the noise on the obtained result. Sub-sampling the raw range data leads to a low image processing overheads like those of segmentation process. Based on interpolation properties of particular wavelets named coiflets, we propose a method for smoothing noisy range images. The smoothed image keeps invariant the "topological characteristics" of the represented surfaces. Thereafter, we propose a method for sub-sampling dense range images which leads to the reduction of the amount of raw data by a factor of four. This method eliminates the "redundant" information, thus the obtained result describes the essential details (as the shape of the physical surface) of the initial range image. The smoothing and sub-sampling methods are designed to be easily integrated in any reconstruction algorithm to improve its result and reduce its overhead in spite of its high complexity.

Collaborative Virtual Environment (CVE) is a promising technology which provides an online shared virtual world to the geographically dispersed people to interact with each other. However, the scalability of existing CVE system is limited due to the constraints in processing power and network speed of each participating host. Large scale CVE (LCVE) that supports a large number of concurrent participants and a large amount of evolving virtual entities is not easy to achieve. In this paper, an autonomous decentralized grid based mobile agent framework for large scale CVE — MACVE is proposed. In MACVE, LCVE software system is decomposed into a group of collaborative mobile agents, each of which is responsible for an independent system task. Agents can migrate or clone dynamically at any suitable participating host (grid node) include traditional servers and qualified user hosts to avoid bottleneck, which can improve the scalability of LCVE. Therefore, MACVE is a logically multi-server architecture while the server workload is autonomously decentralized. Our system prototype and evaluation has demonstrated the feasibility of the proposed framework.

In this paper, we present a method of representing two-dimensional shapes by symbolic features. A shape is represented in terms of multi-interval valued type features. Fuzzy equilateral triangle membership values are used as features to describe a shape. A similarity measure defined over symbolic features useful for matching and retrieval of shapes is also presented. The proposed method of shape representation and matching is shown to be invariant to image transformations (translation, rotation, reflection and scaling) and robust to minor deformations and occlusions. Several experiments have been conducted to demonstrate the feasibility of the methodology and also to highlight its advantages over existing methodologies.

This paper proposes a new improved mountain clustering technique, which is compared with some of the existing techniques such as K-Means, FCM, EM and Modified Mountain Clustering. The performance of all these clustering techniques towards color image segmentation is compared in terms of cluster entropy as a measure of information and observed via computational complexity. The cluster entropy is heuristically determined, but is found to be effective in forming correct clusters as verified by visual assessment.