Tersurat dan Tersirat

Kehidupan sentiasa dipenuhi hikmah. Terpulang kepada kita untuk melihat secara tersurat atau tersirat. "...tidakkah kamu berfikir?."

Sunday, May 2, 2010

My FYP

Dan seperti yang ditulis sebelum ini, saya akan membentangkan projek saya dalam seminar. ^__^

Assalamualaikum and hellos..
The title of my project is comparion of rank filtering techniques using Java. Here, i'm going to talk about the introduction, problem statement, objectives, scope of the project, a little bit of literature review, the methodology and design of this project. Next, will be the implementation and testing. The last but not least is the conclusion.

Before i go further with the introduction, i'm sure most of you in this room must wondering what is the rank filtering? And of course you know what is Java right? ..

Ladies and gentleman,

The title of my project is actually in the field of Image Processing. What is IP? Though i'm not an image processing student, but doing this project, it introduced me that image processing is a term for a wide range of techniques that exit to manipulate and modify the images either color or grey-scale in a various ways. In IP, it includes the image restoration like in photography, image enhancement like what my project is about and image compression.

In this project, i'm using Java as the programming language. Java has the Java 2D API that allows image to be manipulated by using the packages like BufferedImage and BufferedImageOp. I'm also focus only on the grey-scale image. This type of image is an 8-bit image that only contains 1 color in each bit.Now, rank filtering. It is a grey-scale modification that using mapping equation technique where a list of grey level of the neighbourhood pixels is compiled. Then, one of the pixel value is selected as the new value for the pixels.

There are several rank filtering techniques. They are minimum, maximum, median and range filtering. In minimum filtering, the selection of bottom-ranked grey level is selected. In maximum filtering, the top-ranked grey level. In median filtering, the center-ranked and range filtering is the selection of the difference between maximum and minimum. To make you understand, this illustration will make you understand:

So, my project is to make the comparison of images after the rank filtering techniques are applied. I'm using the histogram to illustrate the comparison. The shape of histogram shows the distribution of grey level of an image whether the images are bright or dark image.

Problem statement, in teaching IP, it would be helpful if some there are tools that can used to visualize the various techniques use. Currently, no available visualize and standalone tool that using java as the language.

So, the objectives are to develop a stand alone tool that illustrates the difference of the rank filtering techniques and to differentiate the grey level or can say as the brightness of images using histogram. The scope is grey-scale image and rank filtering techniques.

There are tools that i've found like Filter Forge, Java Image Editor and RankFilteringTool.java which are similar to my project. Of course FilterForge and Java Image Editor are more complete compared to mine, but these tools don't show the difference between the images. So, i tend to make their weakness as my project.

Ladies and gentlemen,
I had used the Rapid Application Development(RAD) as the methodology in my project. In requirement planning phase, i'd made some interview, discussion and try the 3 previous tools.

In the user design, i'd made the system and interface design with the user. System design, the flowchart is used to show the logic of the system and in interface design shows the GUI needed by the user. Durin the construction, the codes are generated and shows the prototype to user. Changes and refinement are made, iteratively, according to user needs. The cutover phase, is final where no changes will be made and just let the user to implement the tool.

Here, some printscreen to show you how the tool works.


And in testing, shows if user load not a grey-scale image(8-bit).

In conclusion, the difference of rank filtering techniques are successfully shown. But to make the system to be standalone, is not successfully achieved. It still needs the java environment to run it.
The objectives are not fully achieved. =) . That's all...Thank you..

No comments: