Restoration of the Image Signal Lost by Analog to Digital Conversion

A/D转换决定了数字图像的分辨率和位长。在视频信号的情况下，帧速率也由视频重新编码决定。许多数字图像/视频设备，如LCD显示器和等离子显示器，都需要图像加密、帧速率上转换和位长扩展技术。本文研究了图像压缩、帧速率上转换和位长扩展三种图像压缩技术。图像压缩：提出了拉普拉斯金字塔表示(LP)方法，并在现场可编程门阵列(Xilinx Virtex-II XC2V6000)上实现。LP方法由5654个切片、7715个FF和42个块RAM实现。帧速率上转换：提出了一种基于块匹配算法的运动补偿的帧速率上转换方法。The Pro…BMA的更多部分随块的大小而变化。对于较小移动对象来说，希望将块大小设置得较小。另一方面，我们希望在全局运动区域中设置较大的块大小。因此，块的大小取决于帧图像局部区域的性质。提出了一种新的基于多尺寸块运动补偿内插的FRUC技术。首先，将帧图像分成块，并估计每个块的运动矢量。接下来，将相似的运动块合并为一个块，最大为32x32个块。位长扩展：提出了一种从M级灰度扩展到N(=3(M-1)+1)级灰度的位长扩展方法。此操作意味着在现有级别之间引入两个灰度级。位长展开法导致了M级灰度图像和N级灰度图像之间的误差图像的估计问题。误差图像仅由三个值构成。如果现有级别和插入的新级别之间的差被定义为Δ，则通过将+Δ、0或-Δ添加到M级别图像来获得N级别图像。大量的应用结果表明，该方法能有效去除数字图像中的伪轮廓。较少

Resolution and bit-length of digital images are decided by Analog to Digital Conversion. In the case of video signals, frame rate is also decided by video recoding. Image enragement, frame rate up-conversion and bit-length expansion techniques are necessary for many digital image/video equipments, such as high-quality display such as the LCD display and the plasma display. In this research, three techniques(i.e., image enragement, frame rate up-conversion and bit-length expansion) are studied.Image enragement : The Laplacian pyramid representation(LP) method, which has been proposed by ourselves, is implemented on FPGA(Xilinx Virtex-II XC2V6000). The LP method is realized by 5654 slices, 7715 FFs and 42 block RAMs. We achieve 206.9[MHz] for maximum operating frequency and only 20 lines for system latency.Frame rate up-conversion : In this research, a novel frame rate up-conversion(FRUC) method using the motion compensation(MC) based on block matching algorithm(BMA) is proposed. The pro … More perty of the BMA is changed by the size of block. It is desirable for small moving object to set block size small. On the other hand, we would like to set the block size large in the global motion region. Therefore, the block size is decided depend on the property of the local region of frame image. A novel FRUC technique based on the motion compensated interpolation with multi-size blocks is proposed. At first, the frame image is divided into 8x8 blocks and motion vectors are estimated for each 8x8 block. Next, the similar motion blocks are merged into one block, up to 32x32 block. Thus the proposed method reduces block artifacts and realizes the clear interpolation of moving objects regardless of the size of those.Bit-length expansion : In this research, we propose a bit-length expansion method from M gray levels to N(=3(M-1) +1) gray levels. This operation means that two gray-levels are introduced between existing levels. The bit-length expansion method results in the estimation problem of an error image between an M gray level's image and an N gray level's image. The error image is constructed by only three values. If the difference between the existing level and the inserted new level is defined as Δ, the N level image is obtained by adding+Δ, 0 or-Δ to the M level image. We show that the proposed method can remove pseudo-contour of digital images through a lot of application results. Less