Image and Video Component Quality Measurements, Analyses, and Applications.
Gao, Chen.
2017-04-20T13:57:38.945Z
-
Abstract: Currently, human experts provide the most reliable image quality evaluations for biomedical, industrial, and security applications. However, subjective evaluation by humans is impractical for real time applications. Therefore, it is desired to create an evaluation process that can be automated and not involve human intervention. This dissertation introduces a novel system for evaluating ... read moreimage quality objectively. The system utilizes multiple image measurements that evaluate one attribute aspect of an image. Each of these attribute measures is inspired by human visual system properties, and can be used separately as a standalone measure for evaluating a certain characteristic of an image. Multiple attributes are then fused together to provide an overall comprehensive evaluation for given image processing tasks. For the purpose of color image and video quality evaluation, attributes such as color, sharpness, and contrast are considered, while in applications where edge map qualities are important, edge pixel localization, corner presence, and double edge occurrence are used instead. To examine the uses of the measures for new image processing algorithm design and analysis, a set of alpha weighted quadratic filter based methods are developed for both color image enhancement and edge detection applications. The presented measures are used to automatically evaluate the image processing algorithm performances, as well as to assist in the selection of optimal operating parameters, all in accordance with the human visual system. The proposed measures are also shown to be applicable as a new method for fast database searching and retrieval. Furthermore, other new applications for the measures for monitoring thermal electrical system conditions, evaluating food quality, inspecting and exploring underwater environment, and for diagnosis and detecting of anomalies in biomedical images are presented.
Thesis (Ph.D.)--Tufts University, 2015.
Submitted to the Dept. of Electrical Engineering.
Advisor: Karen Panetta.
Committee: Sos Agaian, Ronald Lasser, and Ethan Danahy.
Keyword: Electrical engineering.read less - ID:
- 5x21tt20z
- Component ID:
- tufts:21422
- To Cite:
- DCA Citation Guide EndNote
