Use this URL to cite or link to this record in EThOS:
Title: Automatic Aesthetic Image Enhancement for Consumer Digital Photographs
Author: Payne, Andrew
Awarding Body: Loughborough University
Current Institution: Loughborough University
Date of Award: 2007
Availability of Full Text:
Access from EThOS:
Automatic image enhancement refers to the process of improving the quality of the visual content of an image without user interaction. There has been considerable research done in the area of image enhancement normally as a preprocessing step for computer vision applications. Throughout the literature, objective image quality metrics have been defined and image enhancements have been made to satisfy the quality metric. Quality metrics typically are based upon the signal to noise ratio, focus measurements, or strength of edges within the image content. Subjective human input is rarely considered in image enhancement applications. This thesis investigates the concept of automatic image enhancement. In this thesis, an automatic subjective image enhancement system based on the regional content of the image is proposed. The system makes use of segmentation, region classification, image operator discovery and image reconstruction techniques. A detailed description is given of ~ach element of the proposed method and novel methods are proposed in areas where possible improvements can be made to the existing methods with respect to this domain. It is shown in each case that the proposed methods perform well on our J database of images. Overall results of the proposed system are explored with respect to subjective user evaluations. They demonstrate the effectiveness of the overall image enhancement system and possible correlations between the users' feedback are explored. From the analysis, we conclude that the proposed system effectively enhances consumer photographs aesthetically based on the content of the image.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available