Catálogo de publicaciones - libros

In this paper, we present a realistic avatar called e-Shadow. This avatar is a virtual shadow of a user and facial animations are injected into the head position as response to the user’s gestures. We adopt background subtraction method to separate the user’s shape from the background scene and combine several image processing methods in order to reduce noise. A heuristic algorithm for tracking positions of head and hands is proposed to detect predefined user’s gestures. Because e-Shadow assumes everyday indoor lighting, casual camera, and small desktop, it can be directly incorporated with video chatting programs as a realistic avatar and the presented detection method might be applied to the applications of real-time interactive media art.

In this paper, a novel image discriminant analysis method, coined two-dimensional discriminant transform based on scatter difference criterion (2DSDD), is developed for image representation. The proposed 2DSDD scheme adopts the difference of both between-class scatter and within-class scatter as discriminant criterion. In this way, the small sample size problem usually occurred in the traditional Fisher discriminant analysis (LDA) is essentially avoided. In addition, the developed method directly depends on image matrices. That is to say, it is not necessary to convert the image matrix into high-dimensional image vector like those conventional linear discriminant methods prior to feature extraction so that much computational time would be saved. Finally, the experimental results on the ORL face database indicate that the proposed method outperforms Fisherfaces, the standard scatter difference discriminant analysis, not only in the computation efficiency, but also in its recognition performance.

Human motion analysis is an important research subject in human-robot interaction (HRI). However, before analyzing the human motion, silhouette of human body should be extracted from sequential images obtained by CCD camera. The intelligent robot system requires more robust silhouette extraction method because it has internal vibration and low resolution. In this paper, we discuss the hybrid silhouette extraction method for detecting and tracking the human motion. The proposed method is to combine and optimize the temporal and spatial gradient information. Also, we propose some compensation methods so as not to miss silhouette information due to poor images. Finally, we have shown the effectiveness and feasibility of the proposed method through some experiments.

A novel image registration scheme is proposed. In the proposed scheme, two-dimensional principal component analysis (2DPCA) combined with principal component analysis (PCA) is used to extract features from the image sets and these features are fed into feedforward neural networks to provide translation, rotation and scaling parameters. Comparison experiments between 2DPCA combined with PCA based method and the other two former methods: discrete cosine transform (DCT) and Zernike moment, are performed. The results indicate that the proposed scheme is both accurate and remarkably robust to noise.

This paper proposes a novel filtering method for removing such speckle noise from Synthetic Aperture Radar image that combines the Stationary Tree-structured Wavelet Transform (STWT) with a Bayesian wavelet estimator. Experimental results on several test images by using the proposed method show that, the proposed method yields significantly superior image quality and better Peak Signal to Noise Ratio (PSNR).

A new method of the Road extraction in synthetic aperture radar (SAR) image is proposed in this paper. Roads in a high resolution SAR image can be modeled as a homogeneous dark area bounded by two parallel lines. The fundamental is based on linear extraction and improved Hough transform. Combining Hough transform and the partition of the line, the problem of location of line segments was solved. The improved Hough transform is tested on synthetic images and noisy images. The experimental results show the method’s validity.

We consider the generalized regularization problem of Shape-from- Shading. The traditional algorithms are to find the minimum point of the optimization problem where the regularization term is considered as the part of the objective function. However, the result of regularization may deviate from the true surface, due to the ill-posedness of the SFS problem. In this paper, we propose a novel method to solve this problem. The algorithm consists of two steps. In the first step, we recover the components of the surface in the range space of the transpose of the system matrix, from the observed image by using the Landweber iteration method, where the Pentland’s linear SFS model is adopted without any regularization. In the second step, we represent the regularization condition as an energy spline in the Fourier domain, and find the minimum value of the energy function with respect to the components of the surface in the null space of the system matrix. Quantitative and visual comparisons, using simulated data of a fractal and smooth surface, show that the proposed method significantly outperforms the Horn, Zheng-Chellappa, Tsai-Shah and Pentland linear methods for surface reconstruction.

Stack filters are often employed for suppressing the pulse noise. In general, the larger sizes the stack filters are, the better results are. Unfortunately, available algorithms for designing stack filters can only be suit for small window sizes due to their huge computational complexities. This paper presents a new fast adaptive algorithm for designing a stack filter with large windows. The idea of the new algorithm is to divide a lager window into many sub-windows. The procedures of dividing a large window are given. An Immune Memory Clonal Selection Algorithm is employed to design the stack filters with small window sizes. Because of its highly parallel structure, it can be very fast implemented. As an experiment, the algorithm was used to restore images corrupted by uncorrelated additive noise with the level from 10% to 50 %. The results show that the algorithm is effective and feasible.

Cervical cancer is the second most common cancer among women worldwide. Early detection of cervical cancer is very important for successful treatment and increasing survival. We report a spectral imaging microscopic system for Papanicolaou smear analysis for early detection of cervical cancer. Different from traditional color imaging method, we use spectral imaging techniques for image acquisition, which can simultaneously record spectral and spatial information of a sample. In this paper, the imaging instrument construction and spectral image acquisition method is introduced. In the image segmentation process, an effective algorithm using spectral ratio method is applied for cell nuclei detection, which can easily detect the nuclei and diminish the influence of the cytoplasm overlap. Results showed that our segmentation is robust and precise. In addition to this, the segmentation speed is very high.

Recently the method restoring an old picture using the local differential equation on the basis of the geometric measure is proposed. It is necessary to restore an old film and a medical image with noise as well as an old picture. So we extend the method applying for the two dimentional image such as a picture to the three dimensional image such as a time sequence image and a medical image. It is necessary to obtain an object boundary from the original image in order to generate a composite image of good quality, which is difficult to distinguish from the original image. If an object boundary can not be detected, it is difficult to remove the object. In this study, we propose a method for detecting an object boundary and removing it and inpainting its image in a manner that makes it difficult to distinguish from the original image. We extend the image partition method based on the level set method to the method applying for the movie and the medical image to detect an object boundary. We demonstrate its effectiveness by removing a terop from a movie and a tumor from a three dimensional medical image.