- Video Coding and Compression Technologies
- Advanced Vision and Imaging
- Advanced Image Processing Techniques
- Advanced Steganography and Watermarking Techniques
- Chaos-based Image/Signal Encryption
- Advanced Data Compression Techniques
- Digital Media Forensic Detection
- Image and Signal Denoising Methods
- Image and Video Quality Assessment
- Speech and Audio Processing
- Music and Audio Processing
- Video Analysis and Summarization
- Sparse and Compressive Sensing Techniques
- Speech Recognition and Synthesis
- Robotics and Sensor-Based Localization
- Advanced Image Fusion Techniques
- Image Retrieval and Classification Techniques
- Wireless Communication Security Techniques
- Advanced Image and Video Retrieval Techniques
- Advanced Wireless Network Optimization
- Color Science and Applications
- Image Enhancement Techniques
- Mobile and Web Applications
- Indoor and Outdoor Localization Technologies
- Music Technology and Sound Studies
TFI Digital Media Limited (China)
2018-2019
Hong Kong University of Science and Technology
1999-2009
University of Hong Kong
1999-2009
Iowa State University
2009
China Mobile (China)
2009
Education University of Hong Kong
2008
Chinese University of Hong Kong
2005
DSO National Laboratories
2001
Swansea University
1993
University of Toronto
1993
Three novel blind watermarking techniques are proposed to embed watermarks into digital images for different purposes. The designed be decoded or detected without the original images. first one, called single watermark embedding (SWE), is used a bit sequence using two secret keys. second technique, multiple (MWE), extends SWE simultaneously in same space while minimizing (distortion) energy. third iterative (IWE), embeds JPEG-compressed approach of IWE can prevent potential removal JPEG...
In this paper, we propose a coding tree unit (CTU)-level rate control scheme from the perspective of SSIM-based rate-distortion optimization to improve efficiency. First, establish model based on divisive normalization scheme, which characterizes relationship between local visual quality and bits. Then, established is applied CTU-level transformed into global problem solved by convex optimization. Finally, new parameter updating strategy for presented that robust scene variations. Our...
Deep convolutional neural networks (CNNs) have been successfully applied on no-reference image quality assessment (NR-IQA) with respect to human perception. Most of these methods deal small patches and use the average score test for predicting whole quality. We discovered that from homogenous regions are unreliable both network training final estimation. In addition, complex structures much higher chances achieving better prediction. Based findings, we enhanced conventional CNN-based NR-IQA...
Different from the conventional calculative methods, a learning-based initial quantization parameter (LIQP) method is proposed in this paper to improve rate control of high efficiency video coding (H.265). First, framework for (QP) learning proposed, where novel equivalent approach build benchmark labels using single rate-distortion (R-D) pair each QP testing. With criterion maximizing prediction accuracy QPs, features and parameters model are refined. Instead traditionally used target bits...
Rate-distortion (RD)-based mode selections are important techniques in video coding. In these methods, an encoder may compute the RD costs for all possible coding modes, and select one which achieves best trade-off between encoding rate compression distortion. Previous papers have demonstrated that RD-based can lead to significant improvements efficiency. selections, however, would incur considerable increases complexity, since methods require computing numerous candidate modes. this paper,...
High dynamic range (HDR) video compression technology, which is capable of delivering a wider luminance and larger colour gamut than standard (SDR) has been widely used in recent years many fields, including industrial image processing, digital entertainment, machine vision. Rate control (RC) paramount importance to HDR transmission; accordingly, an RC scheme for Efficiency Video Coding (HEVC) proposed this paper. First, considering the characteristics, we propose HDR-Visual Difference...
Motion estimation using the one-bit transform (1BT) was proposed by Natarajan, Bhaskaran and Konstantinides (see ibid., vol.7, p.702-06, 1997) to achieve large computation reduction. However, it degrades predicted image almost 1 dB as compared with full search. We propose a modification 1BT adding conditional local searches. Simulation results show that improves peak signal-to-noise ratio (PSNR) significantly at expense of slightly increased computational complexity. A variant called...
In JPEG-to-JPEG image watermarking (J2J), the input is a JPEG file. After watermark embedding, JPEG-compressed such that output file also We use human visual system (HVS) model to estimate J2J data hiding capacity of images, or maximum number bits can be embedded in images. A.B. Watson's HVS (Proc. SPIE Human Vision, Visual Process., and Digital Display IV, p.202-16, 1993) modified just noticeable difference (JND) for DCT coefficients. The modifications coefficients limited by JND order...
In this paper, we propose a Graphics Processing Unit (GPU)-based motion estimation for H.264/AVC by rearranging the encoding order of 4x4 blocks. Previous research [3] [4] assume no dependency among adjacent blocks, which is not true H.264/AVC. The proposed method only overcomes problem, but also can be used both full search and fast search. Experimental results show that our implementation forty five faster than SDVTD optimized CPU implementation.
Bit-depth expansion is important for displaying a low bit-depth image in high monitor. Existing methods tend to give disturbing contouring or blurring artifacts. In this paper, we propose novel, simple and efficient adaptive method increase taking advantage of the existing techniques superior quality.
Deep learning based image hashing methods learn hash codes by using powerful feature extractors and nonlinear transformations to achieve highly efficient retrieval. For most end-to-end deep methods, the supervised process relies on pair-wise or triplet-wise information provide an internal relationship of similarity data. However, use triplet loss function is limited not only expensive training costs but also quantization errors. In this paper, we propose a novel semantic method for retrieval...
JPEG is a common image format in the world wide web. JPEG-compressed images can be used to hide data for secret internet communication and simply any auxiliary data. In this paper, we propose an algorithm called J-Mark embed invisible watermark information into compressed compress domain. There are three parts of J-Mark: block selection, DCT coefficient modification selected coefficients. Only texture blocks with significant masking properties selection. coefficients energy selected. The...
The presence of noise can dramatically affect the efficiency video compression systems. For performance improvement, most practical systems adopt a denoising filter as pre-processing module for encoder, or post-processing decoder, but complexity introduced by be very high. This paper first presents recursive temporal linear minimum mean squared error (LMMSE) denoising. Based on analysis hybrid process, two novel schemes are presented, one encoding and other decoding, in which proposed LMMSE...
The JVT/H.264 is the new video coding standard which achieves higher efficiency than previous standards such as H.263 or MPEG-2. improvement in rate-distortion performance H.264 mainly due to accurate motion estimation with different mode decision including multiple block size and reference frame estimation. However, full exhaustive search of all sizes computational intensive complexity increasing linearly number allowed size. In this paper, a fast multi-block (FMBME) proposed for coding....
This paper studies the joint security and performance enhancement of secure arithmetic coding (AC) for digital rights management applications. The proposed cryptosystem incorporates interval splitting AC with a simple bit-wise XOR operation step. Security analysis results show that scheme provides satisfactory level against cipher-only attack, chosen-plaintext attack chosen-ciphertext attack. Due to elimination input symbol-wise permutation step, our can be extended conveniently any...
A proper theoretical quantization-distortion model for hybrid video coding is always desirable, since this allows us to explain the behavior of existing codecs and design better ones. However, due existence motion-compensated prediction, introduces interframe dependency into encoded video, which makes its characteristics difficult analyze. In paper, a joint analysis quantization prediction presented. For complete analysis, we investigate not only distortion that signal, but also effect on...
We propose a blind watermarking technique to embed watermark in the JPEG compressed domain. Low frequency DCT coefficients are extracted form an M-dimensional vector. Watermarking is achieved by modifying this vector order point centroid of particular cell. This cell determined according vector, private keys and watermark. A dual-key system used reduce chance removal An iterative approach prevent re-quantization. Experimental results show that can be detected when watermarked image further...
JPEG is a common image format in the WWW and can potentially be used to hide data for secure internet communication watermark copyright control. In this paper, we propose an algorithm embed secret or information. The proposed algorithm, named Watermarking by DC Coefficients Modification (WDCM), assumes that quality factor compression known. We observe it perceptually undetectable if coefficients certain texture-rich blocks are modified small amount. thus information as binary bit sequence...
<para xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <emphasis emphasistype="italic">Multiple description coding</emphasis> (MDC) is an effective emphasistype="italic">error resilience</emphasis> (ER) technique for video coding. In case of frame loss, concealment</emphasis> (EC) techniques can be used in MDC to reconstruct the lost frame, with error, from which subsequent frames decoded directly. With such direct decoding, will gradually recover...
Abstract This article provides two ‘portraits’ of headteachers in primary schools England and Hong Kong, derived from semi-structured interviews with these individuals. Contrary to some claims that such a small sample is worthless, this argues what most meaningful sometimes the singular unique; generalisations education are as likely be useful if they ‘fuzzy’ scientific or statistical generalisations; individual case can strip away clutter large contexts allow recognition common shared...
We propose a novel fast inverse halftoning technique using combination of spatial varying filtering and invariant filtering. The proposed algorithm is significantly simpler than most existing algorithms. Without explicit edge region classification, the filter shown to be capable preserving edges effectively.
In this paper, we propose to embed the watermark information in log-2-spatio domain by means of spread spectrum technique. domain, variance is reduced significantly. This improves efficiency and robustness Low intensity mid-band regions are selected order guarantee an invisible as well JPEG compression. Simulation results show that embedded still survives up compression ratio 14.7.