MEASURING IMAGE QUALITY

PROJECT INFORMATION

Format: ms word /  Chapters: 1-5 /  Pages: 56 /  Attributes: primary data analysis, questionnaires

CHAPTER ONE

1.0      INTRODUCTION

1.1      BACKGROUND OF STUDY

The task of recognizing a person’s face from an frontal image is very easy to human eyes though it is difficult for a machine to perform the same task. In fact it’s a very wide study to perfectly detect a human face (skin) which is very much difficult for a machine. Now there are several types of color spaces like RGB, YCrCb, HSV etc. Now in this paper we use YCrCb color space for two reasons-First, The chrominance information for modeling human skin color can be achieved this color space. Second, for all kind of images, coding is used in YCrCb color format. Now we know that Cr and Cb are the blue and red difference of chroma component. It is proved by experiments that the range of skin color of a human being is always between 133-173 and 77- 127 respectively.

Now the darkness and fairness of human skin is always dependent on ‘Y’ which is the luma component of the color image .Actually ‘Y’ is the brightness of an image. Now after recognizing the face of images especially for video calling purpose or live transmission of Index(Q),Structural Similarity Index(SSIM) and Gradient-based Structural Similarity Index(G-SSIM)news reading images, compression is very much required to get the fast transmission. When an original image file is compressed in other web images format like FLV,3GP,MP4,MPEG and DIVX, the compression techniques compress the images of the video files. Our objective is to show that in all the images compression techniques the face region is less compressed according to the other regions of the image. Visually human eyes can’t detect compression ratio of the face and the other regions of the image.

So to prove this mathematically we have used Universal Image Quality Index (Q), Structural Similarity Index (SSIM) and Gradient-based Structural Similarity Index (G-SSIM). There are other areas of imaging such as security, reconnaissance, and medical imaging, where the image quality focus is on detection and recognition. Detection and recognition typically require a different approach to image quality. Medical image quality, for example, is often evaluated with a different set of tools, such as signal detection theory. Users of these systems nonetheless express their personal preferences, so “beauty contests” still come into play in the design and use of these systems. There have been other proposed definitions for image quality. Janssen and Blommaert have suggested “the quality of an image to be the degree to which the image is both useful and natural.

 The usefulness of an image [is defined] to be the precision of the visual representation of the image, and the naturalness of an image [is defined] as the degree of correspondence between the visual representation of the image and knowledge of reality as stored in memory. This concept assigns to image quality two perceptual attributes: usefulness and naturalness. To use such a definition unnecessarily restricts the concept of image quality to two perceptual attributes or dimensions that may, in fact, be functions of other perceptual dimensions. With the Janssen and Blommaert definition, it is unclear where synthetic or abstract images fit in. For example, how does one characterize the quality of a synthetic image that is not at all “natural,” but useful? Keelan takes a different tack and proposes a definition that considers the image making . “The quality of an image is defined to be an impression of its merit or excellence, as perceived by an observer neither associated with the act of photography, nor closely involved with the subject matter depicted.

 According to this definition, image quality is not, apparently, in the eye of the photographer, the art director, the advertising executive, the producer, director, or a “soccer mom,” to name just a few. Although Keelan makes an interesting case for his definition, the requirements that the observer be distant from the imaging or image making industry and not be involved with the subject of the image are needless and unrealistic complications.

 Although the factors mentioned in some of the above definitions do indeed affect the judgment and preferences of image quality on an individual basis, the view taken with the Image Quality Circle is that these are factors to be controlled or understood in any image quality judgment situation.

STATEMENT OF THE PROBLEM

 Measuring  image quality  plays a vital role in the efficiency of a variety of image processing applications, e.g., image enhancement  image adaptation  image compression, medical imaging, image-based medical diagnosis. Thereof, for the sake of automation and optimization, it highly demands a consistent image quality metric. For instance, in medical applications, the higher enhancement of an image leads to the

earlier and more proper diagnosis of a disease, as well as more successful dealing with epidemics like breast cancer which even in developed countries, one in ten women faces its risk of mortality. In most of the medical images, the indicative features are very small, the detection and interpretation are a difficult task even for an expert physician, and an efficient image enhancement technique helps a lot to avoid improper diagnosis by being used in computer-aided analysis programs. Clearly, the efficiency, optimization, and automation of an image enhancement technique are highly dependent on the deployed metric of image quality.

In recent research works on image quality analysis, there is a major emphasis on a numerical understanding of the human visual system (HVS) to incorporate HVS preferences in image quality indexes. However, it is extremely complex to comprehend HVS preference direction by the current psychophysical tools, but as it is reported even implementation of a simple model of HVS in the quality indexes leads to a higher match with the human observer image quality check. Finally several research has been carried out on Image Quality Assessment for Performance Evaluation of Focus Measure Operators. But not even a single research has been carried out on measuring image quality.

1.2      AIMS AND OBJECTIVES OF STUDY

The main aim of the study is to determine measuring image quality. Other specific objectives of the  study includes;

1.  to determine the relationship on measuring image quality.

2.  to determine the effect of measuring image quality.

3.  to determine the factors affecting measuring image quality.

4.  to determine the extent to which measuring has affected image quality.

5.  to proffer possible solutions to problems.

1.3     RESEARCH QUESTIONS

1. What is the relationship on measuring image quality?

2. What is the effect of measuring image quality?

3. What are   the factors affecting measuring image quality?

4. What is the extent to which measuring has affected image quality?

5.What are the  possible solutions to problems?

1.5      STATEMENT OF RESEARCH HYPOTHESIS

H0: There is no significant difference between measuring image quality.

H1: There is a significant difference between measuring image quality.

1.6      SIGNIFICANCE OF STUDY

The study on measuring image quality will be of immense importance to the entire public, especially those working under image quality organisations, radiographers and film producers, how images should be measured in terms of brightness, contrast in other to have a nice images. It will also enlighten the public the reasons why low image quality should be discourage, because this will lead to low income in an organisation.

It will also enlighten the public the need why good quality of images should be encouraged during the production processes. Finally the study will contribute to the existing literature and knowledge to this field of study and basis for further research.

1.7      SCOPE OF STUDY

The study on measuring is limited to image quality.

1.8     LIMITATION OF STUDY

Financial constraint- Insufficient fund tends to impede the efficiency of the researcher in sourcing for the relevant materials, literature or information and in the process of data collection (internet, questionnaire and interview).

Time constraint- The researcher will simultaneously engage in this study with other academic work. This consequently will cut down on the time devoted for the research work.

1.8      DEFINITION OF TERMS

Image Quality

Can be refer to the level of accuracy in which different imaging systems capture  the qualityof a test and referencimage based on a comparison of features.

Measuring

 Is the assignment of a number to a characteristic of an object or event.

 Quality

 Is define as the measure of excellence. Quality word is used in our daily life like image quality, picture quality, video quality, color quality, etc.