UC知道翻译 好了又追加
来源：百度知道 编辑：UC知道 时间：2024/05/25 01:10:50
不要机器翻译
In rest of the section,we use symbols Ng,Ns and bpp in the figures included in this section to represent size of the gain codebook,size of the shape codebook and bits per pixel,respectively.Wavelet based compression schemes,in general,use higher levels of wavelet decomposition usually in the range of 3–5 levels of decomposition for achieving higher compression ratio.We analyzed the performance of our scheme with 2–4 levels of wavelet decomposition.In general,in all the experiments that we conducted,we observed that the degradation of the SNR is not significant while moving from one level to the next higher level since wavelet transform concentrates maximum signal content into the lower frequency components.However,the bpp decreases significantly allowing the distribution of gained bpp to the detail images.Hence,the SNR does not degrade due to higher levels of wavelet decomposition,if the gained bpp is optimally allocated to the detail images.
For an n-level decompositi
In rest of the section,we use symbols Ng,Ns and bpp in the figures included in this section to represent size of the gain codebook,size of the shape codebook and bits per pixel,respectively.Wavelet based compression schemes,in general,use higher levels of wavelet decomposition usually in the range of 3–5 levels of decomposition for achieving higher compression ratio.We analyzed the performance of our scheme with 2–4 levels of wavelet decomposition.In general,in all the experiments that we conducted,we observed that the degradation of the SNR is not significant while moving from one level to the next higher level since wavelet transform concentrates maximum signal content into the lower frequency components.However,the bpp decreases significantly allowing the distribution of gained bpp to the detail images.Hence,the SNR does not degrade due to higher levels of wavelet decomposition,if the gained bpp is optimally allocated to the detail images.
For an n-level decompositi
在剩下的部分中,我们用符号,坏死性涎腺化生和bpp在这个数字包括在这部分代表大小的改进,得到的大小和形状改进位色,分别是. 基于小波变换的压缩方案,一般情况下,使用高水平的小波分解通常在3 - 5个层次的分解为实现压缩比高我们对我们的方案进行了分析,并对其性能和2 - 4个层次的小波分解一般来说,在所有的实验,我们进行了,我们观察到的降解,无明显的信噪比,同时从一个水平更高水平最高的信号,利用小波变换集中到较低的频率成分含量…然而,bpp明显下降使分布的具体形象推广了因此,信噪比不降低由于更高层次上的小波分解,得到最佳分配给了眩晕是详细的图片。
对于一个n能级分解,我们花了block-size 4×4的地方,HHi sub-image≤我≤1因为较低氮信号的内容,选择总能做的block-size之间,改进大小我们使用相同的数字,因为他们都在为LLn首先对原始图像,这是8个良性阵发性位置性眩晕,利用图像,为国内的一些data-see 9 bpp在Figs.2 bit-allocation子带地图,3 . 在比特率是表示每个像素,应该除以乐队获得位色/乐队良性阵发性位置性眩晕的信噪比,对应的情形0.125代表所有的双正交子类除低频成分被丢弃在4个一级(图4)。小波分解的比较,我们也同时在所有的乐队编码算法的方式使用JPEG-2000我们比较了JPEG-2000信噪比与我们的方案,从四个层面的运用小波分解(虽然不是一个权威的度量的信噪比的评价,在光谱质量2.3节)。