In lossy image/video encoding, there is a compromise between the number of bits (rate) and the extent of distortion. Bits need to be properly allocated to different sources, such as frames and macro blocks (MBs). Since the human eyes are more sensitive to the difference than the absolute value of signals, the MINMAX criterion suggests to minimizing the maximum distortion of the sources to limit quality fluctuation. There are many works aimed to such constant quality encoding, however, almost all of them focus on the frame layer bit allocation, and use PSNR as the quality index. We suggest that the bit allocation for MBs should also be constrained in the constant quality, and furthermore, perceptual quality indices should be used instead of PSNR. Based on this idea, we propose a multi-pass block-layer bit allocation scheme for quality constrained encoding. The experimental results show that the proposed method can achieve much better encoding performance. Keywords: Bit allocation, block-layer, perceptual quality, constant quality, quality constrained
We recently proposed a natural scene statistics based image quality assessment (IQA) metric named STAIND, which
extracts nearly independent components from natural image, i.e., the divisive normalization transform (DNT)
coefficients, and evaluates perceptual quality of distortion image by measuring the degree of dependency between
neighboring DNT coefficients. To improve the performance of STAIND, its feature selection strategy is thoroughly
analyzed in this paper.
The basic neighbor relationships in STAIND include scale, orientation and space. By analyzing the joint histograms of
different neighborships and comparing the IQA model performances of diverse feature combination schemes on the
publicly available databases such as LIVE, CSIQ and TID2008, we draw the following conclusions: 1) Spatial neighbor
relationship contributes most to the model design, scale neighborship takes second place, and orientation neighbors
might introduce negative effects; 2) In spatial domain, second order spatial neighbors are beneficial supplements to first
order spatial neighbors; 3) The combined neighborship between the scales, spaces and the introduced spatial parents is
very efficient for blind IQA metrics design.
This paper presents a statistical reconstruction algorithm for dual-energy (DE) CT of polychromatic x-ray source. Each
pixel in the imaged object is assumed to be composed of two basis materials (i.e., bone and soft tissue) and a penalizedlikelihood
objective function is developed to determine the densities of the two basis materials. Two penalty terms are
used respectively to penalize the bone density difference and the soft tissue density difference in neighboring pixels. A
gradient ascent algorithm for monochromatic objective function is modified to maximize the polychromatic penalizedlikelihood
objective function using the convexity technique. In order to reduce computation consumption, the
denominator of the update step is pre-calculated with reasonable approximation replacements. Ordered-subsets method is
applied to speed up the iteration. Computer simulation is implemented to evaluate the penalized-likelihood algorithm.
The results indicate that this statistical method yields the best quality image among the tested methods and has a good
noise property even in a lower photon count.
Dual-energy X-ray imaging is an important method of medical imaging, capable of not only obtaining spatial information of imaging object but also disclosing its chemical components, and has many applications in clinic. The current computation methods of dual-energy imaging are still based on the model of mono-energy spectrum imaging with some linear calibration, while they are incapable to reflect correctly the physical characteristics of dual-energy imaging and obstruct deeper research in this field. The article presents a new medical X-ray imaging model in accordance with physics of imaging and its corresponding computational method. The computation process includes two steps: first, to compute two attenuation parameters that have clear physical meaning: equivalent electron density and attenuation parameter of photoemission; then to compute the components of high- and low-density mass through a group of simple equation with two variables. Experiments showed that such method has quite a satisfactory precision in theory, that is, the solutions of parameters under different exposure voltages and thickness of tissue for several main tissues of human body are much low in deviations, whose quotient of standard deviation divided by mean are mostly under 0.1%, and at most 0.32%. The method provides not only a new computational way for dual-energy X-ray imaging, but also a feasible analysis for its nature. In addition, the method can be used to linearly rectify data of dual-energy CT and analyze the chemical component of reconstructed object by means of parameters clear in physics.
A novel method of the matching and reconstruction of DSA vessel axis is proposed based on the redundant information from multiple two-dimensional (2-D) projections of the object. First, a correlation scheme on pixels gray level is used to extract the vessel structure on every projection. Through this way, we acquire a binary image. Secondly, Hit-Miss Transform (HMT) is applied on the binary image to produce a vessel axis image. Thirdly, an arbitrary projection is chosen as base image, and others are chosen as reference images. For each key point in the base image, the matching points are found from the key points of the reference images according to epipolar geometry and the topological linking relations of vessel branches. The matching segments are determined from matched key points. If one segment of vessel axis in the base image is confirmed to be matching with one in the reference image, an interpolation process may be used to find out the corresponding relationship of points on these two segments. Then we can get a continuous 3-D segment of DSA vessel axis according to reconstruction the matched points between two views. The whole process is executed repeatedly when another image is selected as base image. The experiment result shows that this method may provide satisfactory reconstruction even the vessel extraction result is not very accurate.
The three-dimensional reconstruction of human brain vessels has been the research focus in recent year. Reconstructing the axis and the cross section of vessels respectively is one effective method in 3D vessel reconstruction. However the conventional reconstruction methods usually take the complicated space relations of different projections into counts to match a vessel segment in different angles. Obviously, this kind of method is very convoluted and can not be easily used in general situations. Considering that the underlying cause of the difficulty in reconstruction is the complexity of brain vesselsí» intricate directions and bifurcation styles, this paper proposes a novel method of vessel axis reconstruction based on vessel bifurcation model. The experiments are satisfying.