The application of image enhancement in color and grayscale images
Presented By: Nisar Ahmed Waqas Ahmed HITEC University Taxila Department of Electrical Engineering
Image enhancement techniques are used to emphasize and sharpen image features such as to obtain a visually more pleasant, more detailed, or less noisy output image. Image enhancement is the process of applying these techniques to facilitate the development of a solution to a Digital imaging problem.
The aim of image enhancement is to improve the interpretability or perception of information in image for human viewers, or to provide `better input for other automated image processing techniques. Application specific image enhancement
Contrast Enhancement Brightness Contrast Linear Contrast Stretching Histogram Equalization Adaptive Contrast Enhancement Color Contrast Linear Color Contrast Color Balance Blur Reduction Image Sharpening Weiner Deconvolution Removing Noise Linear Filtering Median filter Adaptive filtering
Histogram of images enhanced by Linear Stretching, Histogram Equalization and Local Contrast Enhancement.
The images in figure 1 shows the Images captured in low light areresult of color contrast darker and a low contrast. In theenhancement, the image on the figure 2 the image on the left is theleft is a washed out image having original darker image which is firstpoor contrast. After the application transformed into a lighter image byof color correction technique shifting histogram towards rightwhich works separately on the RGB side and then applying colorlayers the resultant image has correction by converting it intomuch better detail and looks like a intensity image.good image for visual perception.
Image captured in ambient light Weiner deconvolution filterhaving a color other than white provide us best results in debluringmay have a color cast. An aerial when we know the length andimage may also have a color cast angle of distortion in motiondue to low quality LSR or some blurred images. The resultsother reason. One such image is become more and more accurateshown in figure 3. This effect can when we put a value close to theeasily be eliminated by applying original distortion. Figure 5 showshistogram equalization on the result of motion blurred imageintensity layer. restored by using correct length and angle of distortion.
Cameras having autofocus take a little time to focus the subject so prior capturing ofimage may hay some lens blur. If the camera is in high f-stop it is tuned to capturenear object and blur the far objects. If a landscape image is captured in high f-stopsettings the whole image has a Gaussian blur. We can remove this by using Weinerfilter by adjusting its parameters to Gaussian filter which sharpen the imagefeatures. The image in figure 4 is blurred by this effect and has a slightly lowcontrast. This image is corrected by sharpening the image followed by histogramequalization.
The Salt and Pepper type noise is typically caused by malfunctioning of the pixelelements in the camera sensors, faulty memory locations, or timing errors in thedigitization process. For the images corrupted by Salt and Pepper noise, the noisypixels can take only the maximum and the minimum values in the dynamic range.Figure 6 shows an image corrupted by salt and pepper noise. The image is restoredby using median filter. The image on the right side is restored image but its edgesand sharpness is blurred due to median filtering. The blurring can be increased ordecreased by varying the radius of median filter.
Sparkle noise causes artificial artifacts in digital images. A bright spot exist in imagewhich have a typical intensity of 40%. This effect can easily be reduced by usingsigma filter by adjusting the value of threshold. Figure 7 shows the image corruptedby sparkle noise and then restored by adjusting a suitable value of sigma filter. Thisimage shows that sigma filter produce much better results than median filter insome cases.
Histogram equalization shows best result in most of the cases but if theimage has a wide light color area we will incorporate adaptive histogram equalization.
Sigma filter is best among all the discussed techniques because it betterpreserve the edges while removing the noise, we can use it as a medianfilter by adjusting the value of threshold to 100%.
In image sharpening and motion blur reduction only twotechniques are discussed which produces best results in theirapplication area.
Image enhancement techniques are widely available, but their applications are not well defined. Application software has been designed to check the effect on the filter before its application. A detailed discussion has been made on the base of results to select an algorithm on the base of filtering requirements. These techniques are tested on a large number of images and have shown significant results.
various aspects of image enhancement are catered for in the implementation and subsequent exercise of results, nevertheless, we understand that it is so demanding and absorbing area for research that the work could substantially be carried forward in following directions as a future work: Improvement in selective noise reduction techniques. Level correction of image by combining it with image segmentation. Noise reduction by anisotropic diffusion using closed edges.