Posture is an action expression of the specific requirements of the limbs and body interaction or human behavior. But human movement often has a self-occlusions or sheltered phenomenon, such as the hands swing and foot alternating when a man walking. For body occlusion problem, based on human perception, humans have the repairing ability to automatically represent body appearance and predict action mode. So far, there is not a powerful technology to solve this problem, but it is very important and valuable scheme especially environment security surveillance or patient attention. Hence, in this paper, an efficient approach for repairing the lost body information is proposed. The processing steps of this approach consist of two phases: training and testing. The training phase is to capture human skeleton information via shooting the depth image based on natural interaction mechanism and to construct the skeleton database. After training, the correction parameters and scene setting can be obtained to use in testing phase. The testing phase is to reconstruct the lost skeleton of human under shelter and to evaluate the system performance. Experimental results show that the rebuilt human skeleton is very close to the real skeleton like unsheltered situation.

In this paper, a lossless data-hiding approach is presented based on quantized coefficients of discrete wavelet transform (DWT) in the frequency domain to embed secret message. Using the quantized coefficients for 9/7 wavelet filter in DWT, we embed secret data into the successive zero coefficients of the medium-high frequency components in each reconstructed block for 3-level 2-D DWT of a cover-image. The procedures of the proposed system include embedment, extraction, and restoration. Experimental results show that the proposed method can achieve high embedding capacity and acceptable image quality of stego-image, and data reversibility.

In this paper, a fast motion estimation algorithm combined hybrid predictor and search pattern (called HPS) is proposed to find the best motion vector (MV). Block-based motion estimation (ME) and pattern searching strategy are used to detect the fast vector motion. The advantages of the approach can improve the performance of existing fast ME algorithm and can efficiently use to video compression. In addition, we also adopt the early termination strategy into this search method to speed up on searching motion vectors (MVs). The experimental results demonstrate that the proposed algorithm is faster than diamond search (DS) and hexagon-based search (HS), and the image quality, on average, is raised to 0.2dB superior to those of search method, and can effectively improve the reconstructed frame quality and be suitable for real-time applications.

The increasing popularity of the internet means that digital multimedia are transmitted more rapidly and easily. And people are very aware for media ownership. However, digital watermarking is an efficient and promising means to protect intellectual properties. Based on the intellectual property attention in the information era, how to protect the personal ownership is extremely important and a necessary scheme. In this paper, we propose an effective video watermarking method based on a pseudo-3-D discrete cosine transform (DCT) and quantization index modulation (QIM) against several attacks. The watermark is mainly inserted into the uncompressed domain by adjusting the correlation between DCT coefficients of the selected blocks, and the watermark extraction is blind. This approach consists of a pseudo-3-D DCT, watermark embedding, and extraction. A pseudo-3-D DCT, which is taken DCT transformation twice, will be first utilized to calculate the embedding factor and to obtain the useful messages. Using the QIM, we embed the watermark into the quantization regions from the successive raw frames in the uncompressed domain and record the relative information to create a secret embedding key. This secret embedding key will further apply to extraction. Experimental results demonstrate that the proposed method can survive filtering, compressions, luminance change, and noise attacks with a good invisibility and robustness.

In this paper, we formulate the timing-driven multiple-layer obstacle-avoiding rectilinear Steiner minimal tree (TML-OARSMT) problem and propose an algorithm to solve the problem. The objective is simultaneously to minimize the maximum source-to-terminal length and minimize the number of vias. The proposed algorithm contains the three steps, including the multiple-level timing-driven partitioning, the multiple-level routing tree construction with obstacles, and the timing-driven multiple-layer balanced tree construction. Compared with the traditional method, the partition-based method can significantly improve the maximum source-to-terminal length and the total wirelength by 11.3% and 10.9% respectively with the additional number of vias.

In this paper, we propose an approach to categorize the film classes by using low-level features and visual features. The goal of this approach is to classify the films into genres. Our current domain of study is using the movie preview. A film preview often emphasizes the theme of a film and hence provides suitable information for classification process. In our approach, we classify films into three broad categories: action, dramas, and thriller films. Four computable video features (average shot length, color variance, motion content and lighting key) and visual effects are combined in our approach to provide the advantage information to demonstrate the movie category. Our approach can also be extended for other potential applications, including the browsing and retrieval of videos on the Internet, video-on-demand, and video libraries.

There is a substantial interest in locating bookbacks in a bookrack image using the edge information contained in the image. The effective edge information considered here is defined between two adjacent bookbacks, called a "bookback edge". The main steps of the proposed approach are as follows: the edge points of the input image are first detected by a simple thresholding method, a measure algorithm applied to the edge information locate all possible bookback edges, and the final bookback edges are obtained by averaging the fuzzy bookback edges. The experiments show that the proposed approach is feasible.