Real-Time Imaging最新文献

Complexity localisation in the reference frame is the key process for the derivation of efficient scanning orders for motion estimation. The more localised the complexity is, the more computationally efficient scanning orders can be derived for reduced cost motion estimation algorithms. However, this processes entails serious pre-processing overhead which may render it unsuitable for real time video coding systems. In this paper, we propose three low complexity scanning orders of similar performance that are very competitive in terms of the operation count ratio metric with respect to the MPEG-2 raster scan order, show improvements of 7.14% on the average with respect to the number of examined macroblock rows metric and they also show an increase in the speed-up ratio of 0.12 on the average as compared to the standard. As compared to other work in the literature, the proposed scanning orders require one fourth of the operation count ratio and show an increase in the speed-up ratio of 45 times on the average.

This paper proposes two frame-based adaptive thresholding algorithms for reducing the amount of computation involved in block-based motion estimation in the real-time and low-rate video coders. The proposed algorithms use an adaptive threshold for each of the frames to prejudge if each of the block matching computation is worthwhile. Based on the difference between the current and reference frames, the first algorithm determines a threshold in terms of the optimal trade-off between run-time and distortion, and the second algorithm determines a threshold according to a user specified percentage of total number of blocks. The experimental results demonstrate that the proposed algorithms can significantly reduce the amount of computation compared to the previous algorithms, while almost fully maintaining the quality of the reconstructed image.

This paper presents a new hardware design for a neural network based colour image compression. The compressed image consists of a colour palette containing few best colours and the coded image. Kohonen's map neural network is applied to construct the colour palette and the coded image, both forming the compressed image. The Kohonen's map based compression results in linear time complexity (in the size of the image). It is advantageous over traditional JPEG in colour quantization applications and compression of images with limited colours. The architecture of the hardware unit is based on single instruction multiple data methodology. The architecture has been implemented in an application specific integrated circuit and results show that the proposed design achieves high speed allowing inputs at a video rate for compression of images up to size of 512×512 with low area requirement.

A color correction method for balancing the color appearances, among a group of images, about a specified object or scene, such as panoramic images and object movies, is developed and tested in this paper. In order to increase the running speed of color correction and reduce the out-of-gamut color pixels, we introduce the selection of principal regions. The average color values of principal regions are used to construct the low-degree (up to degree two) polynomial mapping functions from the source images to the corrected images. The functions are run in the decorrelated color spaces. Our method is tested using real and synthetic images. The results of these tests show the proposed method can get a better performance than other existing methods.

This paper presents two robust algorithms with respect to global contrast changes: one detects changes; and the other detects stationary people or objects in image sequences obtained via a fixed camera. The first one is based on a level set representation of images and exploits their suitable properties under image contrast variation. The second makes use of the first, at different time scales, to allow discriminating between the scene background, the moving parts and stationarities. This latter algorithm is justified by and tested in real-life situations; the detection of abnormal stationarities in public transit settings, e.g. subway corridors, will be presented herein with assessments carried out on a large number of real-life situations.

This paper deals with parallelization and implementation aspects of partial differential equation (PDE)-based image processing models for large cluster environments with distributed memory. As an example we focus on nonlinear diffusion filtering which we discretize by means of an additive operator splitting (AOS). We start by decomposing the algorithm into small modules that shall be parallelized separately. For this purpose image partitioning strategies are discussed and their impact on the communication pattern and volume is analyzed. Based on the results we develop an algorithmic implementation with excellent scaling properties on massively connected low-latency networks. Test runs on two different high-end Myrinet clusters yield almost linear speedup factors up to 209 for 256 processors. This results in typical denoising times of $\text{0.4}\text{s}$ for five iterations on a 256×256×128 data cube.

Although the hardware platform is often seen as the most important element of real-time imaging systems, software optimization can also provide remarkable reduction of overall computational costs. The recommended code development flow for digital signal processors based on the TMS320C6000^TM architecture usually involves three phases: development of C code, refinement of C code, and programming linear assembly code. Each step requires a different level of knowledge of processor internals. The developer is not directly involved in the automatic scheduling process. In some cases, however, this may result in unacceptable code performance. A better solution can be achieved by scheduling the assembly code by hand. Unfortunately, scheduling of software pipelines by hand not only requires expert skills but is also time consuming, and moreover, prone to errors. To overcome these drawbacks we have designed an innovative development tool—the Software Pipeline Optimization Tool (SPOT^TM). The SPOT is based on visualization of the scheduled assembly code by a two-dimensional interactive schedule editor, which is equipped with feedback mechanisms deduced from analysis of data dependencies and resource allocation conflicts. The paper addresses optimization techniques available by the application of the SPOT. Furthermore, the benefit of the SPOT is documented by more than 20 optimized image processing algorithms.