There are two main methods in digital image enhancement: spatial domain method
and frequency domain method. Spatial domain method mainly processes the pixel in
the image on the basis of gray mapping transformation. Frequency domain method
is based on convolution theorem. In general, frequency domain method uses
frequency transform such as Fourier transform method to achieve image

Frequency domain method

       Gamma correction
Gamma correction is a spatial domain technique. As is well known, the quality of the
image could be influenced by many factors, such as the weather, the precision of the
camera, optical system, etc. The images captured might be lighter or darker than
expected. For this kind of quality degradation, gray level stretching can be used to
achieve the aim. By redistributing the original image’s gray levels, an image could
have better quality.
Input image

1.2.2. Histogram equalization
The principle of histogram equalization is to distribute the gray-scale equalization of
the original image uniformly from a typical gray zone to the entire gray range.
Histogram equalization is a nonlinear stretching of the image. It re-allocates image
pixel values and makes the amount of the pixels in every certain gray range as
similar as possible. Histogram equalization changes the histogram distribution of a
given image to a uniform histogram distribution.
Histogram equalization sometimes would reduce gray-scale so that some details
would disappear after converting.
1.2.3. Discrete Cosine Transform (DCT)
Besides image enhancement in the spatial domain, there are also many image
enhancement techniques performed in frequency domain, such as the image
enhancement based on Fourier transform, discrete cosine transform (DCT), and so
on. The advantage of image enhancement in frequency domain lies in that image
enhancement can be performed by selecting the frequencies. This is very important
for low-vision patients because some patients cannot see some special frequencies.
One important image enhancement in the frequency domain is DCT-based image
enhancement techniques. Besides the advantage mentioned above, DCT-based image
enhancement techniques can save time when the image to be enhanced in JPEG
format because we can perform the enhancement in the decoding stage.
The DCT based image enhancement for a JPEG image in the decoding stage can be
described by Figure 2 [14].

Figure 2. Image enhancement in JPEG domain based on the method described in
1.2.4. Image enhancement with wavelet transform
Besides image enhancement techniques in spatial and frequency domains, another
image enhancement technique is performed in a time (spatial)-frequency domain.

Wavelet transform-based image enhancement is one of these techniques. In recent
decades, wavelet transform has become an important topic in image processing areas
and has found wide applications in different areas. It has been applied to image
compression, noise reduction, image enhancement and so on.
In wavelet transform based image enhancement [15], the image is first transformed
into wavelet domain. The image enhancement can be achieved by the modification
of the wavelet coefficients. Different techniques to modify the wavelet coefficient
techniques have been proposed. One of the techniques is the technology developed
by Dr. Jinshan Tang in [16]. The basic idea of Tang’s algorithm is to define a multicontrast measure in the wavelet domain. The enhancement can be achieved by
directly modifying the contrast of the image. The proposed method has been found
very effective for mammogram enhancement.
The advantage of wavelet transform-based image enhancement lies in the
decomposition of the content of the original image into multi-scale frequencies
which match the human visual system and thus the enhancement can be performed
using human visual features. Another benefit lies in that the enhancement can save
time when the images to be enhanced are in the JPEG2000 format.
1.3 Statement of Problem
The aim of this thesis is to investigate new image enhancement algorithms and their
application in mobile device (iPhone). The proposed algorithm is based on wavelet
transform and visual weighting. In the proposed algorithm, the key point is the
contrast enhancement in wavelet domain through contrast sensitive function masking.
Experiments through changing the contrast manipulation factors were performed to

get the best enhanced images. In the experiments, subject tests were used to test the
performance of the algorithms by comparing images enhanced by the proposed
algorithm and by Tang’s algorithm. Besides, a mobile application for visually
impaired was developed. The application could run on IOS devices such as iPhone,
iPad. Users with an iPhone installed with the proposed application can use the
iPhone as a test tool to get the enhanced version of an image.
The rest of the thesis is organized as follows. In Chapter 2, some basic wavelet
transform knowledge and Tang’s algorithm are introduced. Then, I investigate a
visual weighting image enhancement algorithm based on human visual system. In
Chapter 3, IOS system is introduced. On the IOS platform, Tang’s algorithm and
visual weighting image enhancement algorithm are implemented. Chapter 4
describes the experiments which are used to compare the two algorithms.