Reconfigurable computing techniques and VLSI implementations for IT security and image processing

可重构计算旨在填补硬件和软件之间的差距，潜在地实现比软件高得多的性能，同时保持比硬件更高水平的灵活性。通过将应用程序的计算密集部分放置到可重构硬件上，它将大大提高系统的性能。在无线设备和嵌入式系统上实现用于IT（信息技术）安全的密码服务提出了巨大的挑战，特别是对存储器、功率、计算资源和移动性的限制。可重构计算可以作为一个很好的平台，在有限的计算环境和无线设备上的混合密码系统。可重构体系结构提高了系统方案的灵活性和计算资源的利用率，降低了系统总成本。 此外，硬件盗版问题导致IT行业利润损失巨大。 在本研究计画中，我们也探讨如何使用可重组态计算/架构来保护硬体的知识产权。 图像处理在医学、犯罪预防和调查、出版、娱乐、教育和科学研究等领域起着至关重要的作用。 快速、自动地从图像数据库中找到所需的图像已成为当务之急。这种技术被称为基于内容的图像检索（CBIR）。然而，很少有人研究使用可重构计算技术来提高图像检索的速度。我们将探索新的图像检索方法，并提出有效的算法和快速实现，以提高目前的CBIR系统的检索速度。我的研究目标是研究和设计低成本、高速度、高灵活性的可重构体系结构，以支持实时在线安全数据通信、硬件IP（知识产权）保护和图像检索。所提出的可重构结构具有规则的结构，非常适合于高速VLSI实现。

Reconfigurable computing is intended to fill in the gap between hardware and software, potentially achieving a much higher performance than software, while maintaining a higher level of flexibility than hardware. By placing the computational-intense portions of an application onto the reconfigurable hardware, it will accelerate the system performance greatly. The implementations of cryptography services for IT (Information Technology) security on wireless devices and embedded systems present a great challenge, particularly to the memory, power, computation resources, and mobility constraints. Reconfigurable computing can serve as an excellent platform for hybrid cryptosystems on restricted computing environments and wireless devices. The reconfigurable architectures increase the flexibility of system schemes and the efficiency of computation resources, and reduce the total cost. Moreover, hardware piracy issues cause huge profit loss in the IT industry. In this research project, we also examine how to use the reconfigurable computing/architecture to protect the hardware IP (Intellectual Property). Image processing plays a crucial role in areas of medicine, crime prevention and investigation, publishing, entertainment, education, and scientific research. It has become urgent to find the desired images from an image database quickly and automatically. This technique is known as content-based image retrieval (CBIR). However, little research has been done to use the reconfigurable computing technique to speed-up the image retrieval. We will explore new approaches of image retrieval and propose efficient algorithms and fast implementation to improve the retrieval rate of current CBIR systems. My research goal is to investigate and design reconfigurable architectures with low costs, high speeds, and high flexible structures to support real-time online secure data communications, hardware IP (Intellectual Property) protection, and image retrieval. The proposed reconfigurable architectures have regular structures and are very suitable for high speed VLSI implementation.