Computational methods for image processing understanding and recognition

图像处理理解和识别的计算方法

• 批准号: 9265-2010
• 负责人: Bui, Tien
• 金额: $ 1.82万
• 依托单位: Concordia University
• 依托单位国家: 加拿大
• 项目类别: Discovery Grants Program - Individual
• 财政年份: 2013
• 资助国家: 加拿大
• 起止时间: 2013-01-01 至 2014-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=521247
• 关键词: Computational; methods; image; processing; understanding

In today's world there is an increasing demand for digital image processing and understanding for the purpose of automation of information systems such as extraction, classification, search, and retrieval. In the sub-discipline of digital document processing, the amount of paper documents that must be processed by human in many organizations both commercial and governmental offices for archival purposes is huge and growing every day. There is an urgent need for automation of this process. For clean printed documents, the problem could be considered as already solved at least in theory. However, the difficulty is in dealing with unstructured, unconstrained handwritten documents that could be subjected to degradation over time or noise and artifacts due to the scanning process. Thus, the research on document image processing and understanding spans a wide range of sub-disciplines of computer science including image processing, pattern recognition, natural language processing, machine learning, database systems, and information retrieval. The objective of this research is to study the general problem of computational methods for digital image processing and understanding. This includes digital document processing as a major component as well as medical imaging, biometrics, and related topics. New generation of techniques for computational image processing should combine new development in image modeling with intelligent approaches based on human vision principle, learning based methods, and pattern recognition techniques to develop intelligent tools. Recently, computational intelligence techniques such as neural networks or evolutionary algorithms have been employed in various applications in the area of medical imaging. Approaches based on computational intelligence have been shown to be advantageous compared to classical approaches. Examples where this research would be useful are in document image processing, medical imaging, security, and biometric applications.

在当今世界，为了信息系统的自动化，例如提取、分类、搜索和检索，对数字图像处理和理解的需求日益增加。在数字文档处理的子学科中，在许多商业和政府办公室中必须由人类处理的用于存档目的的纸质文档的量是巨大的，并且每天都在增长。迫切需要实现这一过程的自动化。对于干净的打印文件，至少在理论上可以认为这个问题已经解决。然而，困难在于处理非结构化、无约束的手写文档，这些文档可能会随着时间的推移而退化，或者由于扫描过程而产生噪声和伪影。因此，对文档图像处理和理解的研究跨越了计算机科学的广泛子学科，包括图像处理、模式识别、自然语言处理、机器学习、数据库系统和信息检索。本研究的目的是研究数字图像处理和理解的计算方法的一般问题。这包括作为主要组成部分的数字文档处理，以及医学成像，生物识别和相关主题。新一代的计算图像处理技术应该将图像建模的新发展与基于人类视觉原理的智能方法、基于学习的方法和模式识别技术相结合，开发智能工具。近来，诸如神经网络或进化算法的计算智能技术已经被用于医学成像领域中的各种应用中。与经典方法相比，基于计算智能的方法已被证明是有利的。这项研究将是有用的例子是在文档图像处理，医学成像，安全和生物识别应用。