量化步长

现有量化步长估计算法是基于DCT系数的统计直方图进行。

Current quantization step estimation algorithm is based on the histogram of DCT coefficients .

一种估计JPEG双重压缩原始量化步长的新方法

A New Method for Estimating the Primary Quantization Step of JPEG Double-Compression

在新的调节器中，限制了量化步长（Q）的范围，消除单个或者少数几帧产生的编码误差。

In new adjuster the range of quantization step has been limited well and coding error has been removed in single or some little frames .

在传统R-D模型中，码率R和失真D都看成量化步长的函数，而且失真评价函数一般在象素域或空域中进行。

In conventional R-D analysis , the bitrate R and distortion D are considered as function of quantization parameter q , which are usually developed in pixel-domain or spatial domain .

由于JM模型采用线性模型预测平均绝对差值（MAD）以获得模式选择所需要的量化步长，因此算法运算量较大，而且MAD预测准确度十分有限。

As JM uses linear model in MAD prediction to achieve magnitude step , the calculation cost of this algorithm is fairly high and the precision of MAD prediction is limited .

阐述了采用两个步骤设计H.263量化器：首先利用缓冲器纯占有控制算法来确定量化器的粗量化步长；

In this paper how to design a quantizer of H.263 through two steps is introduced : First , obtaining the coarse quantize step through buffer pure occupy algorithm ;

在为DWT系数确定好一组使给定码率下均方误差（MSE）最小化的最佳量化步长{qi}后，该量化器将能够实现小波系数的动态量化。

The quantizer can perform dynamic quantization for wavelet coefficients after the determination of the best quantization steps { qi } that minimize the average mean square error at a given bit rate .

分析和实验结果表明，水印信息嵌入在量化步长较小的量化DCT系数的最低位，既不影响算法抗JPEG压缩的能力，又能有效缓解待保护DCT系数个数与不可见性之间的矛盾。

The simulation results indicate that embedding watermark in the LSB ( Least Significant Bit ) of the DCT coefficients corresponding to smaller quantization step in the quantization table can remit the conflict between the DCT coefficients to be protected and invisibility without dropping the capability of resisting JPEG compression .

同时，根据ICM中最大值出现的位置能定位出图像的块边缘。其次，本文改进了现有量化步长的估计算法，并提出了两种估计量化表的算法。

At the same time , the location of block boundary can be localized through the location of the maximum of the ICM . Secondly , this paper improves the current quantization step estimation algorithm , and proposes two novel algorithms to estimate quantization table .

由于量化步长与乘性噪声能同步变化，因此在原理上该算法对乘性噪声具有良好的鲁棒性。

Since the quantization step changes synchronously with the multiplication noises , this algorithm is robust to multiplication noises in principle .

水印嵌入时利用加密技术来提高水印的安全性，依据小波低频系数自适应的选取量化步长。

Using encryption technology to improve security , it means to adaptive select the quantization step in wavelet low frequency coefficient .

通过分析量化步长与压缩比之间的统计规律，提出基于统计规律的量化步长设计，减少冗余编码。

Through analyzing the statistical law between quantization steps and compression ratios , a quantization step design algorithm is proposed to reduce time overhead .

针对这些缺陷，在小波系数的合理组织、多门限量化电平的设置、扫描次数等方面作了相应的研究，从理论上推导了不同分解尺度之间量化步长的成倍数关系。

To solve those pitfalls , some studies have been done in regard to reasonable organization of wavelet coefficients , multi-threshold setting for quantization and scan times .

提出了一种可用于版权认证的基于小波变换域自适应量化步长的图像盲水印算法。

A new image blind watermarking algorithm based on adaptive quantization step in discrete wavelet transform ( DWT ) is proposed , which can be used in copyright authentication .

而且在用统一的量化步长时，我们分别就嵌入部分分别为图像的某一区域和各个区域进行讨论。

And when the quantization step in a region is the same one , we discussed it when the embedded component is one region of the image or all .

该算法以抖动调制为基础，以视频帧的运动特征和纹理特征作为边信息，自适应地确定抖动调制的量化步长。

In the algorithm , dither modulation is adopted , and the motion feature and texture feature are used as side information to determine the adaptive quantification step of every frame .

根据各个分块的统计特性，利用神经网络来确定的各个分块的量化步长，将水印嵌入到各个分块的奇异值。

The quantization step is adjusted through the neural network according to the statistics of the blocks , and then the watermark is embedded on the singular value of the blocks .

高频线性调频雷达的距离量化步长通常为几公里，相对舰船目标尺度而言测距精度非常低。

HF chirp radars usually have quantized range measurements with a quantization step of several kilometers , which offer too low an accuracy for ranging of targets such as ships and planes .

本文充分利用直方图中量化步长整数倍处其局部极大值的特点，改进了现有算法，使之能更好地估计长量化步长。

This paper exploits the fact that the multiples of the quantization step are local maximum in the DCT histogram , and improves the current algorithm . Improve algorithm can estimate long quantization step well .

本文在仔细分析了编码参数对视频编码影响的基础上，指出量化步长是影响视频编码能耗、编码码率和编码质量的重要参数之一。

After carefully analyzing the effect of quantitative parameters for video coding , the paper points out the influence from the quantization step for video coding energy consumption , video coding rate and the video coding quality .

通过分别估计量化步长估计出的量化表不一定完全正确。本文将该不完整的量化表和所有已知量化表进行匹配，最终得到了完整的量化表。

The estimated quantization table obtained from estimating every single quantization step may contain incorrect estimated quantization steps . The half-baked quantization table is matched with all has-known quantization tables and the best matched quantization table is chose .

因此，在电池能量状态划分的前提下，建立了三种能量级别下的基于量化步长的率失真模型。（3）建立不同能量状态下的手语视频编码功率率失真模型。

Therefore , on the premise of battery energy state division , the rate-distortion model which based on quantization step is established under the three energy levels . ( 3 ) Establish video coding power-rate-distortion model under different energy states .

由于图像层和文本层具有不同的量化方法和量化步长，这里提出了联合质量控制方法来平衡二者的量化误差。

A joint quality control method is proposed to balance the quantization errors of the pictorial layer and the textual layer .

水印的嵌入采用量化调制方式，其量化步长可根据人类视觉系统和载体图像的局部特点自适应地调整。

Watermark embedding is performed by means of quantization index modulation , the quantization step of which can be dynamically adjusted according to the local image characteristics and the human visual system .

由于该方案可以直接确定量化参数，从而避免了由量化步长转化为量化参数产生的舍入误差。

With the established scheme , the quantization parameter can be determined directly , thus avoiding the round-off error , which is brought by the transformation from quantization step into quantization parameter .

因此，选用量化参数作为建模参数，建立了基于量化步长的视频编码能耗模型。

Therefore , the quantitative parameters are chosen as the model parameters and the energy consumption model of video coding is established by quantization step .

在量化过程中，根据人眼视觉模型可见临界误差来确定每个系数的量化步长，实现了水印的自适应嵌入。

Quantization step of each coefficient is automatically computed in terms of the JND to adaptively embed watermarking .

现有量化调制音频水印技术普遍采纳了均匀量化策略，其不仅需要结合大量实验确定量化步长，而且难以保证不可感知性与鲁棒性的良好平衡。

Uniform quantization strategy has been widely adopted by currently available quantization modulation techniques of audio watermarking . It not only needs to ascertain the quantization step with massive experiments , moreover it 's difficulty in guaranteeing the fine balance between the transparency and robustness .

在给定的比特率条件下，通过对使得解码图像量化误差达到最小的最佳比特分配策略进行研究，给出了在均匀量化情况下的改进三维小波变换结构的量化步长。

Then the optimal bit allocation strategy with minimized distortions at a given bit rate is presented and the quantization steps are given for the improved algorithm .