Zhang Yongdong

Fast 4/spl times/4 intra-prediction mode selection for H.264

In this paper, we propose a fast intra-prediction mode selection method for JVT video coding standard H.264. Based on local edge information obtained by calculating edge feature parameters, the method can reduce the computational complexity considerably while maintaining similar PSNR and bit rate. Experimental results are presented to show the effectiveness of the proposed method

Efficient Tm:YAG Ceramic Laser at 2 μm

We demonstrate a high-efficiency continuous-wave Tm:YAG ceramic laser pumped with a Ti:sapphire laser. An output power up to 860 mW is obtained under an absorbed pump power of 2.21 W at 785nm, corresponding to a slope efficiency of 42.1% and optical to optical efficiency of 22%. The measured central wavelength is 2012 nm.

Evaluation of hydrocarbon generation of the Xiamaling Formation shale in Zhangjiakou and its significance to the petroleum geology in North China

The semi-closed pyrolysis simulation system under constant pressure was conducted to explore the characteristics and mechanisms of hydrocarbon generation from Xiamaling Formation shale in Xiahuayuan, North China. The experiment results indicate the oil generated by the Xiamaling Formation shale in oil window should be classified as “aromatic-intermediate” type, whereas the decreasing of dry coefficient can be ascribed to the cracking of residual bitumen in source rock in the stage of high to post maturity. The amount of hydrocarbon gas generated from residual bitumen can be up to 1–2 m3 per ton rock in high to post mature stage by calculating hydrogen contents in the kerogen, the expelled hydrocarbon, and the residual hydrocarbon. This reveals the importance of residual bitumen as a gas source during high to post mature stage of the kerogen evolution, and also as the broad exploration prospect of these gases. This research highlights the attention should be paid to oil/gas reservoirs sourced from residual bitumen of organic-rich source rock in high mature stage, even the primary oil/gas reservoirs considered as the main exploration targets in middle-upper Proterozoic sediments of North China.

Adult Image Detection Combining BoVW Based on Region of Interest and Color Moments

To prevent pornography from spreading on the Internet effectively, we propose a novel method of adult image detection which combines bag-ofvisual-words (BoVW) based on region of interest (ROI) and color moments (CM). The goal of BoVW is to automatically mine the local patterns of adult contents, called visual words. The usual BoVW method clusters visual words from the patches in the whole image and adopts the weighting schemes of hard assignment. However, there are many background noises in the whole image and soft-weighting scheme is better than hard assignment. Therefore, we propose the method of BoVW based on ROI, which includes two perspectives. Firstly, we propose to create visual words in ROI for adult image detection. The representative power of visual words can be improved because the patches in ROI are more indicative to adult contents than those in the whole image. Secondly, soft-weighting scheme is adopted to detect adult images. Moreover, CM is selected by evaluating some commonly-used global features to be combined with BoVW based on ROI. The experiments and the comparison with the stateof-the-art methods show that our method is able to remarkably improve the performance of adult image detection.To prevent pornography from spreading on the Internet effectively, we propose a novel method of adult image detection which combines bag-of-visual-words (BoVW) based on region of interest (ROI) and color moments (CM). The goal of BoVW is to automatically mine the local patterns of adult contents, called visual words. The usual BoVW method clusters visual words from the patches in the whole image and adopts the weighting schemes of hard assignment. However, there are many background noises in the whole image and soft-weighting scheme is better than hard assignment. Therefore, we propose the method of BoVW based on ROI, which includes two perspectives. Firstly, we propose to create visual words in ROI for adult image detection. The representative power of visual words can be improved because the patches in ROI are more indicative to adult contents than those in the whole image. Secondly, soft-weighting scheme is adopted to detect adult images. Moreover, CM is selected by evaluating some commonly-used global features to be combined with BoVW based on ROI. The experiments and the comparison with the state-of-the-art methods show that our method is able to remarkably improve the performance of adult image detection.

Panoramic video coding using affine motion compensated prediction

This paper proposes an affine motion compensated prediction (AMCP) method to predict the complex changes between the successive frames in panoramic video coding. A panoramic video is an image-based rendering (IBR) technique [1] which provides users with a large field of view (e.g. 360 degree) on surrounding dynamic scenes. It includes not only the translational motions but also the non-translational motions, such as zooming and rotation etc. However, the traditional motion compensated prediction is a translational motion compensated prediction (TMCP) which cannot predict nontranslational changes between panoramic images accurately. The AMCP can model the nontranslational motion effects of panoramic video accurately by using six motion coefficients which are estimated by Gauss Newton iterative minimization algorithm [2]. Simulated results show that the gain of coding performance is up to about 1.3 dB when using AMCP compared with TMCP in panoramic video coding.This paper proposes an affine motion compensated prediction (AMCP) method to predict the complex changes between the successive frames in panoramic video coding. A panoramic video is an image-based rendering (IBR) technique [1] which provides users with a large field of view (e.g. 360 degree) on surrounding dynamic scenes. It includes not only the translational motions but also the non-translational motions, such as zooming and rotation etc. However, the traditional motion compensated prediction is a translational motion compensated prediction (TMCP) which cannot predict non-translational changes between panoramic images accurately. The AMCP can model the nontranslational motion effects of panoramic video accurately by using six motion coefficients which are estimated by Gauss Newton iterative minimization algorithm [2]. Simulated results show that the gain of coding performance is up to about 1.3 dB when using AMCP compared with TMCP in panoramic video coding.

Ontology Based Sports Video Annotation and Summary

With digital sports video increasing everyday, effective analyzing sports video content becomes more and more important. Effective and efficient representation of video for searching, retrieval, inference and mining is a key problem in knowledge engineering. To describe sports video content efficiently, sports video ontology for video annotation is represented in OWL, a description logic based Web Ontology Language. We describe a user-friendly platform for sports video annotation. Ontology based sports video annotation can facilitate video indexing, retrieval and reasoning in a broad range of applications including Digital Olympic Project in China. Moreover, we present a hierarchical sports video summarization strategy to browse the sports video in a progressive way. In sports video, replay scenes often represent the highlight or interesting event of the video. Hence, our representative scene selection is based on the replay detection algorithm and identical events detection. The basic experimental results show our strategy is effective.

A Tunable Ultrafast Source by Sum-Frequency Generation between Two Actively Synchronized Ultrafast Lasers

We demonstrate an experimental setup of a tunable ultrafast laser source by sum-frequency generation (SFG) between a mode-locking Ti:sapphire laser and a Nd:YVO4 laser. The generated wavelength by SFG is tunable from 450 nm to 480 nm with timing jitter no more than 1 ps. The average output power is over 20 mW and the maximum is about 30 mW at 457 nm. This ultrafast laser is a simple and easy tuning source applied to some pump-probe spectroscopy and ultrafast dynamics experiments.

A pseudo relevance feedback based cross domain video concept detection

Due to the mismatch of data distribution between training and testing data set, the issue of semantic gap in the field of video concept detection becomes more and more serious. To solve this problem, an effective pseudo relevance feedback (PRF) based method is proposed in this paper to build domain adaptive classifiers. Firstly, the mechanism of PRF tries to select some pseudo samples according to the fused estimation for test data given by existing source models. Then, these pseudo samples are integrated into the process of Tradboost based cross domain transfer learning to make the best use of semantic information generalized by existing source models. Extensive experiments demonstrate that the proposed method can effectively enhance the performance of cross domain learning.Due to the mismatch of data distribution between training and testing data set, the issue of semantic gap in the field of video concept detection becomes more and more serious. To solve this problem, an effective pseudo relevance feedback (PRF) based method is proposed in this paper to build domain adaptive classifiers. Firstly, the mechanism of PRF tries to select some pseudo samples according to the fused estimation for test data given by existing source models. Then, these pseudo samples are integrated into the process of Tradboost based cross domain transfer learning to make the best use of semantic information generalized by existing source models. Extensive experiments demonstrate that the proposed method can effectively enhance the performance of cross domain learning.

Nd:YAG Lasers Operating at 1064nm and 946nm by Direct Pumping and Thermally Boosted Pumping

We demonstrate a 1064nm Nd:YAG laser by directly pumping into the upper lasing level with a tunable Ti:sapphire laser. The valid wavelength is demonstrated at 868.3 nm, 875.2 nm, 883.8 nm, and 885.5 nm, respectively. To our knowledge, this is the first time that 1064 nm Nd:YAG laser pumped by 875.2 nm laser. In addition, laser wavelength at 946 nm is also generated by direct pumping together with traditional pumping.

Highly Efficient Self-Starting Femtosecond Cr:Forsterite Laser

We report a highly efficient and high power self-starting femtosecond Cr:forsterite laser pumped by a 1064-nm Yb doped fibre laser. Five chirped mirrors are used to compensate for the intra-cavity group-delay dispersion, and the mode-locking is initiated by a semiconductor saturable absorber mirror (SESAM). Under pump power of 7.9 W, stable femtosecond laser pulses with average power of 760mW are obtained, yielding a pump power slope efficiency of 12.3%. The measured pulse duration and spectral bandwidth (FWHM) are 46 fs and 45nm; the repetition rate is 82 MHz.