IEEE Trans Image Process. 2014 Oct;23(10):4426-37. doi: 10.1109/TIP.2014.2348869. Epub 2014 Aug 18.
Contrast enhancement is crucial when generating high quality images for image processing applications, such as digital image or video photography, liquid crystal display processing, and medical image analysis. In order to achieve real-time performance for high-definition video applications, it is necessary to design efficient contrast enhancement hardware architecture to meet the needs of real-time processing. In this paper, we propose a novel hardware-oriented contrast enhancement algorithm which can be implemented effectively for hardware design. In order to be considered for hardware implementation, approximation techniques are proposed to reduce these complex computations during performance of the contrast enhancement algorithm. The proposed hardware-oriented contrast enhancement algorithm achieves good image quality by measuring the results of qualitative and quantitative analyzes. To decrease hardware cost and improve hardware utilization for real-time performance, a reduction in circuit area is proposed through use of parameter-controlled reconfigurable architecture. The experiment results show that the proposed hardware-oriented contrast enhancement algorithm can provide an average frame rate of 48.23 frames/s at high definition resolution 1920 × 1080.
在图像处理应用中，生成高质量的图像时，对比度增强至关重要，例如数字图像或视频摄影、液晶显示处理和医学图像分析。为了实现高清视频应用的实时性能，需要设计高效的对比度增强硬件架构以满足实时处理的需求。在本文中，我们提出了一种新颖的面向硬件的对比度增强算法，可有效地应用于硬件设计。为了考虑硬件实现，提出了近似技术来减少对比度增强算法执行过程中的这些复杂计算。所提出的面向硬件的对比度增强算法通过测量定性和定量分析的结果来实现良好的图像质量。为了降低硬件成本并提高实时性能的硬件利用率，通过使用参数控制可重构架构来提出电路面积的减少。实验结果表明，所提出的面向硬件的对比度增强算法可以在高清晰度分辨率 1920×1080 下提供平均 48.23 帧/秒的帧率。
