Catálogo de publicaciones - libros

This paper presents a novel approach to implement estimation and recognition of human motion from uncalibrated monocular video sequences. As it is difficult to find a good motion description for humans, we propose a matching scheme based on a and a , to detect individual body parts and analyze the shape of the whole body as well. In a frame-by-frame process, both descriptors are combined to implement the matching of the motion pattern and the body orientation. Moreover, we have added a novel spatial-temporal cost factor in the matching scheme which aims at increasing the temporal consistency and reliability of the description. We tested the algorithms on the CMU MoBo database with promising results. The method achieves the motion-type recognition and body-orientation classification at the accuracy of 95% and 98%, respectively. The system can be utilized for an effective human-motion analysis from a monocular video.

Feature extraction is one of the most important problems in face recognition task. In this paper, we use kernel uncorrelated discriminant analysis to extract the optimal discriminant features for face recognition. The method also solves the so-called “Small Sample Size” (SSS) problem, which exists in most Face Recognition tasks. Experimental results on the Yale face database and AT&T face database show the effectiveness of this method.

A novel method for automatic occluded targets recognition in SAR images is proposed in this paper. Different SAR occluded targets are simulated based on actual vehicles from the MSTAR database, and are recognized using SVM classifier by grouping recognition based on the targets azimuth angles. It is shown that the proposed method outperforms the typical methods in accuracy at high occlusion, and robustness to occlusion with experiments considering accuracy and confusion matrix.

Image collections may contain multiple copies, versions, and fragments of the same image. Storage or retrieval of such duplicates and near-duplicates may be unnecessary and, in the context of collections derived from the web, their presence may represent infringements of copyright. However, identifying image versions is a challenging problem, as they can be subject to a wide range of digital alterations, and is potentially costly as the number of image pairs to be considered is quadratic in collection size. In this paper, we propose a method for finding the pairs of near-duplicates based on manipulation of an image index. Our approach is an adaptation of a robust object recognition technique and a near-duplicate document detection algorithm to this application domain. We show that this method requires only moderate computing resources, and is highly effective at identifying pairs of near-duplicates.

Recently, Thien and Lin proposed a user-friendly (, )-threshold scheme which employs Lagrange interpolation to produce shadow images representing a shrunken version of the original image. In this paper, we introduce a new framework which uses more ( – 1)-degree polynomials with different primes to enhance the functionality of the user-friendly (, )- threshold framework and obtain more effective performance for large k. Since the proposed framework significantly reduces reconstruction errors compared to the previously published user-friendly schemes, it is suitable for modern visual communication applications where features, such as distributed trust, secure transmission and storage, fault tolerance, and high-quality image reconstruction are required.

Recently, various non-photorealistic rendering (NPR) techniques have been developed for computers to generate images of different artistic styles automatically. Due to the complexity of the algorithms, however, most NPR methods are limited to the processing of still images. It is the objective of this paper to extend and improve existing NPR techniques to enable near real-time processing of video. The enhancement will be achieved by compressed domain processing. By exploiting the relationship among I, P, and B frames, different strategies can be developed to increase the efficiency of the NPR algorithm. Experimental results have demonstrated the efficacy of the proposed methods and validated the near real-time creation of NPR video effects.

In this paper, we shall propose an image-hiding scheme to embed color secret images into their compression codes. The color host image is compressed by using the CSMVQ algorithm. Meanwhile, the color secret image is embedded during the production of the compression code. To get this job done, we must have the power to partition the state palette to sub-state palettes quickly and temporarily. In our new method, we sort the different colors in the palette to gather similar colors. In addition, there is a modulo operator we use to partition the state palette dynamically. This way, the secret data can be embedded into the compression code after the state palette is partitioned. According to our experimental results, the image quality of both the color host image with the secret data embedded in and the reconstructed color secret image is good when our proposed method is used.

In this paper, an interface using eye and mouse for the handicapped people is proposed. The eye regions are localized using neural network (NN)-based texture classifier that discriminates the facial region into eye class and non-eye class, and then are tracked using mean-shift procedure. The mouse region is detected based on edge information and then tracked using template matching. To assess the validity of the proposed system, it was applied to the interface system and was tested on a group of 25 users. The results show that our system has the accuracy of 99%.

The appreciation of various mobile multimedia services causes a paradigm shift in the Internet. Since the introduction of the route optimization in the Mobile IPv6, various approaches are proposed to solve the problem such as the handover latency and the excessive signaling message exchanges and enhance security. In spite of this improvement they are not enough for the low-battery mobile devices to service the audio and video streaming seamlessly because of the handover time and the computation of the mobile node. We propose an advanced agent-delegated route optimization protocol for seamless multimedia services at low-battery mobile devices.

Media environment has been rapidly changed these days. Internet provides multimedia environment and has become a part of our everyday life as a new advertisement media. Authentication methods typically involve an authentication server in the home network from the authentication server of the external network, and then the external network is notified with the authentication result in the home authentication server. This paper discusses and proposes a key distribution protocol that ca be used when the user moves the external network and to ensure secure communication with the external authentication server. As the proposed technology does not receive the reissue of authentication from the home authentication server in case moving between the external authentications servers, authentication server overload can be reduced and the home network can be made to issue the key using the registered password.