EXPERT SYSTEM

Digital Image Processing

Digital image processing is a process that aims to manipulate and analyze images with the help of computers. Digital image processing can be categorized into two types of activities:

1. Improving the quality of an image, so it can more easily interpreted by the human eye.
2. Processing information on an image for automatic object recognition.

The second application field is very closely related to science pole (pattern recognition) are generally intended to recognize an object in a way to extract important information contained in an image. When associated with pattern recognition image processing, are expected to form a system that can process the input image so that image pattern recognition. This process is called image recognition or image recognition. Image recognition process is often applied in everyday life.

Image processing and pattern recognition to be part of the image recognition process. Both applications will be complementary to a characteristic of an image that would be recognized. In general, digital image processing steps include image acquisition, image quality, image segmentation, representation and description, recognition and interpretation.

Image acquisition

The data can be done by using various media such as analog cameras, digital cameras, handycamp, scanners, optical readers and so on. The resulting image may not be digital data, so the need didigitalisasi.

Improved image quality

At this stage known as pre-processing where in improving the image quality can increase the likelihood of success on the stage of the next digital image processing.

Image segmentation

Segmentation aims to select and isolate (separate) an object of the whole image. Segmentation consists of downsampling, filtering and edge detection. Downsampling stage is a process to reduce the number of pixels and remove some information from the image. With a fixed image resolution, image size downsampling to produce smaller. The next stage is screening segmentation with median filter, this is done to eliminate noise that usually appears at high frequency in the spectrum image. In screening with the median filter, image gray level at each pixel is replaced with the median value of the pixel gray levels contained in the filter window. The last stage in the process of segmentation is edge detection. Canny algorithm approach is based on the convolution function of the image with Gaussian operator and its descendants. Edge detector is designed to represent an ideal edge, with the desired thickness. In general, the process of segmentation is very important and directly

will determine the accuracy of the system in the iris identification process.

Representation and Description

Representation refers to the data conversion from the segmentation results into a form more suitable for processing on the computer. The first decision that must have produced at this stage is the data to be processed within the boundaries or area complete. Boundary representation is used when the emphasis is on the characteristics of the outer shape, and area of representation is used when the emphasis is on the characteristics of, for example texture. After the data has been represented to form a more appropriate type, the next step is to describe the data.

Introduction and Interpretation

Pattern recognition is not only aiming to get an image with a certain quality, but also to classify the various images. From a number of images processed so that images with similar characteristics are grouped in a particular group. Interpretation of the meaning of emphasis include the recognized object.