Image enhancement techniques and applications

Image Enhancement Techniques

Research shows that out of the five senses which are hearing, smell, sight, touch, and taste -- which humans utilize to observe their environment, sight is the most influential (Jeong, 2011). Analyzing images and getting them really does form a huge part of the unchanging cerebral activity of human beings during the course of their lives. Actually, beyond 98% of the activity of the human brain is included in managing images from the visual cortex (Guruvareddy & Giri Prasad, 2011). In today's communications system it is vital to recognize that the multimedia is an area that is continually increasing.

Basically, it is a field that is growing more and more each day. Many are starting to see the various avenues that a person can go into when it comes to image enhancement techniques. There used to be an era when the options were very limited, but now thanks to that advancement in technology, times have changed and therefore, there is a growing field. Many are not aware that there are actually increasing demands on incorporating visual features to other styles of communications. It is as a result unable to be avoided to have circumstances in which the video and conveyed images being corrupted or degraded in their perceptual quality by diversity of methods. With that said, this essay will discuss the new and different image enhancement techniques.

What is Image Enhancement?

Image enhancement is basically refining the interpretability or insight of data in imageries for human spectators and delivering 'better' input for other automatic image processing type of techniques. The principal goal of image enhancement is to adapt qualities of a copy to make it more suitable for an assumed task and a particular observer. Image enhancement is considered to be among the simplest and most attractive parts of digital image processing. Essentially, the idea that goes on behind enhancement techniques is to be able to bring out a lot of detail that is hidden, or just to highpoint assured features of interest in an image. It is vital to bear in mind that enhancement is a very personal area of image processing (Guruvareddy & Giri Prasad, 2011). Enhancement in quality of these tainted images can be attained by using application of enhancement methods.

During this process, one or more attributes of the image are then able to be modified. The selection of attributes and the way they are adapted are exact to a given task. Furthermore, observer-specific factors, for instance the human visual system and the spectator's experience, will present a great deal of bias into the choice of image enhancement approaches (Jeong, 2011). There exist a lot of techniques that can improve a digital image without having it spoiled. The enhancement approaches can approximately be divided in to the following two groups: Spatial Domain Methods and Frequency Domain Methods.

What are the Techniques?

Research shows that the Image processing techniques are mostly utilized in order to figure out the alignment and range of pictures or imageries. Various enhancement techniques are utilized for enhancing an image which comprises things such as filtering, gray scale manipulation, and Histogram Equalization (HE). Most of these techniques are used to brighten up the image and make it better. As time has gone on, there have been so many different techniques that have worked and several that have not. Even though these methods do preserve the input brightness on the output image with an important contrast enhancement, they could possibly produce images they may or may not look as natural as the input ones (Gorgel, & Ucan, 2010).

Spatial Domain Methods and Frequency Domain Methods

In spatial domain techniques (Gorgel & Ucan, 2010) people directly deal with the image and a lot of pixels. The pixel values are then manipulated in order to achieve the wanted enhancement. In the frequency domain approaches, the image is first transferred in to frequency domain. In Spatial Domain methods normally denotes to the image plane itself. Research shows that image processing methods, spatial domain methods are founded on uninterrupted influence of pixels in a picture. Many are actually unaware that the spatial filtering and intensity transformations are two major groups of spatial domain approaches.

However, when it comes to the Frequency domain methods, first image is mostly converted to frequency domain. This is saying that, the Fourier transform of the image is calculated and then performed by being processing on the Fourier transform of the image. Lastly, Inverse Fourier transform is done in order to receive image that is resultant (Jeong, 2011).

It is clear that every one of the enhancement functions are mostly done on the Fourier transform of the image and also the Inverse Fourier transform is done to get the subsequent image. All of these types of enhancement functions are done in order to adjust the image contras, brightness, or the delivery of the grey levels. When it comes to the result, the pixel value (strengths) of the output image will be adapted as stated by the transformation function operated on the input standards into image T. using G. Wherever T. is the transformation. The standards of pixels in images g and f are denoted by s and r, correspondingly (Guruvareddy & Giri Prasad, 2011). All of this is very important in inaging because without them, there is no detail in the image.

Digital Image Processing

Digital image processing is a method that is hot on the market and being used by many as far as image enhancing. An image is basically described as two- dimensional function, f (x, y). It is also defined where x, y are plane coordinates and the amplitude of 'f' at any pair of coordinates (x, y) is mentioned as the gray or intensity level of the image. However, when x, y and the intensity values of f are all limited and separate quantities, people normally just refer to the image as a typical digital image. When it comes to processing the image by means of computer algorithms, then it goes by the name of digital image processing. When it is compared to analog image processing, digital image processing basically has many advantages. It has these types of benefits because it is able to avoid problems for instance image degradation, signal distortion, and build-up of noise throughout processing.

A Digital Gray Level Image Technique

This is another technique that can be used. However, it is probably not looked at as being the most popular. This image does not really give that much detail and people are mostly interested in features however, this is a digital gray image and is looked at as being a simple two dimensional matrix of figures which basically extends from 0 to255. A lot of these types of numbers usually to have a custom of representing varied shades of gray. However, the number '0' is what shows pure black color and then it will display the number '255' which in presents pure white color (Reddy & Kumar, 2012).

Create Negative of an Image Technique

Another technique involves the negative image. This technique is used by a lot of those that understand the process and the methods of image enhancement. The most basic and easiest operation in digital image processing is when they have to be able to compute the negative of an image. The pixel gray values are then inverted in order to compute the negative of what the image is supposed to be. This is said by many to be the simplest technique because it does not take much effort to actually operate. However, this technique is also the most used partly because it is the easiest. Those that do not want to worry much about the contrast will use the technique.

Figure 1 this is the output and input of negative imagery.

Brightness Control Technique

Another technique is using the brightness control. This is also easy for some to adjust but most may not be interested in this technique because they do not want their image to bright. This technique is very helpful for people because for instance if the digital image is of poor brightness, then of course the objects in the image will not be noticeable obviously. It should be the situations when the image is seized under conditions that are believed have low light conditions (Guruvareddy & Giri Prasad, 2011). If people want to correct this problem, then it is possible that they can further make sure that the brightness of the captured digital image is turned up more and they can also make the image a lot more attractive.

Research shows that if an individual studies the histogram of a low-brightness image, they will be able to discover that the most of the pixels will probably lie in the left half of the gray value range. Most do not realize that the brightness does bring more enhancements as far as imaging. This method is a lot of people's favorite especially for those that are interested in images that are very bright. When it is bright, it gives the picture a lot more contrast and also more details unlike the negative images where there is not much detail.

Many individuals like this technique because the brightness of a dark image can effortlessly be augmented by putting in a constant to gray value of the entire pixel. Research about this enhancement, shows that this operation can possibly shift the histogram in the direction of a brighter side with a factor that is constant. In spite of the fact that applying this technique to increase brightness of an image, an individual will have to select the continuous more wisely in order for the complete range of gray values to lie somewhere within the 0 to 300 range. Most will not want the image if it happens to be more than 300 because the image will not appear as clear (Guruvareddy & Giri Prasad, 2011). Research does show that if the concluding gray worth of any pixel is in excess of 300 then the information could possibly be lost.

Contrast Stretching Technique

Contrast stretching is also a technique that is used vastly. In fact, this has a huge audience of people that like this as well. Research shows that this type of method is much better for the image quality enhancement in contrast to brightness control. Experts on this issue make the point that if low contrast image is caused because of low light circumstances, then the absence of lively range of the camera sensor, contrast stretching operation will have an outcome that will be of good quality image. All through the contrast stretching operation, people will essentially have the dynamic range of the gray values increased. Many have discovered that they are able to utilize a lot of the functions for contrast stretching.

Figure 2 an example of the contrast stretching of color.

Adaptive Histogram Equalization method

During the early 1980s, there had been a lot of work on c.c.d. Here, all of the scanners were reviewed and solid-state scanners were embraced on what were called the-chip signal processing functions (Reddy & Kumar, 2012). Of course during this time imaging had not yet reached its height. Ever since that time, future trends have been leaning towards what are now called 'smart' scanners. These are scanners that involve the histogram.

Histogram is something that is not used that much but some finds it beneficial. Research shows that this is an addition to traditional Histogram Equalization technique. This is sort of well-known because it does things such as enhance the contrast of images by converting the values that are in the intensity part of the image I.

Figure 3 this shows the detail of the histogram.

However, unlike HISTEQ, it has the tendency to function on what is considered to be the small data regions (tiles), instead of the entire image (Reddy & Kumar, 2012). Everyone of the tile's contrast is normally enhanced, in order for the histogram of the output region to roughly match the stated histogram. The adjacent tiles are then jointed utilizing bilinear interpolation so that artificially induced boundaries will be eliminated. The contrast, particularly in homogeneous places, can be restricted so that amplifying the noise which might be present in the image can be avoided.

Dualistic sub-image histogram equalization Technique

By many experts, this is looked at as being the novel histogram equalization technique in which the image that is considered to be the original is putrefied into two equal area sub-images which are based on its gray level likelihood density purpose. Next, the two sub-images are supposed to be equalized correspondingly. In the end, the people are able to get some of the results after the processed sub-images which are normally collected into one appearance. In truth, the algorithm can not just improve the image visual information efficiently, nonetheless also coerce the original appearance's regular luminance from great shift (Gorgel & Ucan, 2010). When this is done, it really does make it much more possible to be used in video system much more directly.

Image enhancement algorithm Technique

The image enhancement algorithm does things like first causing an image to separate into its lows which are known by many as the low-pass filtered form and then there is what is called the highs also recognized as the (high-pass filtered form) components. Experts make the point that the lower component is what controls the amplitude of the highs component in order to raise the local contrast (Jeong, 2011). After that occurs, the lows component is then exposed to a non-linearity to adapt to the local luminance mean of the image and is jointed with the administered highs constituent.

According to Umamaheswari & Radhamani (2012) experts make the point that the Enhancement algorithms are usually based on interquartile distances and local medians which are considered to be more effective than those utilizing standard deviations and means for the elimination of spike noise, and other things that can actually distort the images. However, they are not as fast as the mean-standard deviation counterparts but are appropriate for large data sets preserved in small machines in manufacture quantities (Gorgel, & Ucan, 2010).