Notes on Digital Image Processing - I

(Definition) Image: it's a two-dimentional function $f(x,y)$ where $x$ and $y$ are spatial cartesian coordinates and the intensity of $f$ at any $(x,y)$ is called the intensity or gray level of that image at that point.

(Definition) Digital Image: When $x,y,f$ are all finite, we call the image a digital image.

Digital Image Processing has two main general purposes: improvement of pictorial information for human interpretation, and image processing for storage, transmission and representation for autonomous machine perception.

It's not easy to clearly differentiate the area of digital image processing from the area of computer vision (which simulate the human vision, involving learning and taking actions based on input images), there's a continuum between them. But this can be divided into three intervals, represented by their processes: low-, mid- and high-level processes. The first one encompasses the image preprocessing: noise reduction, contrast enhancement and image sharpening, being both the input and output images. The second one is about segmentation (partitioning of the image into areas and objects), description of those segments in order to make them intelligible for computer processing,  and classification (recognition) of individual objects. The latter one is about assign meaning to combinations of recognized objects.

The overlap area between image processing and image analysis, based on the previous paragraph, is the recognition of indivial regions and objects in an digital image. So, a definition of digital image processing can be stated as:

(Definition) Digital Image Processing: set of processes whose input and output are images and also processes that extract attributes from images, up to and including the recognition of individual objects.

In order to check some of the various applications of this field, it's a good approach to classify images according to it's souce of energy. Below, some application from the eletromagnetic waves are listed:

• Gamma-ray Imaging:  nuclear medicine and astronomical observations;
• X-ray Imaging: medical diagnosis, industry, astronomy;
• Ultraviolet band: lithografy, industrial inspection, microscopy, lasers, biological imaging, astronomical observations;
• Visible and Infrared bands: light microscopy, astronomy, remote sensing, industry, law enforcement;
• Microwave band: radar;
• Radio Band: medicine and astronomy;

Besides the eletromagnetic waves, there are other important sources of energy that can form images, as seen below:

• Acoustic imaging: geological exploration, industry, medicine;
• Electron microscopy:
• Synthetic imaging: fractals, 3-D modeling;

In order to work with digital image processing, it's important to be aware of some fundamental steps, which can be grouped in two categories: the ones related to processes for which the input and the output are images, and the ones for which the output is made of attributes extracted from the images. These steps are:

• Image acquisition: sometimes involves acquiring the image and doing some preprocessing such as scaling;
• Image enhancement: process of manipulating an image so thar the result is more suitable than the original for a specific application;
• Image restoration: a very objective process based on mathematical or probabilistic models of image degradation;
• Color image processing
• Wavelets: foundation for representing images in various degrees of resolution;
• Compression: techniques for reducing the storage required to save an image, or the bandwidth required to transmi it;
• Morphological processing: deals with tools for extracting image components that are useful in the representation and description of shape;
• Segmentation: procedures to partitionate an image into its constituent parts or objects;
• Representation and description: almost always follows the output of the segmentation phase, which is usually raw pixel data; involves deciding if the data should be represented as boundaries or regions; the description (feature selection) deals with extracting attributes that result in some quantitative information of interest or are basic for differentiating one class of object from another;
• Recognition: assigns labels to an object based on its descriptors.

All of these stages are connected to (and by) a knowledge base, which is the knowledge about a problem domain. This base is generally coded into an image processing system in the form of a knowledge database.

It's worth to point that a digital image processing task may not require all of those processes, it all depents of it's complexity.

Digital Image Processing Systems have a set of componentes described below:

• Image sensors: two elements are required, the physical device to capture the energy irradiated by the object we wish to image, and a digitizer for converting the output of the physical sensing device into digital form;
• Specialized image processing hardware: the digitizer plus hardware that performs other primitive operations, such as an arithmetic logic unit (ALU);
• Computer: general purpose computer, can range from a PC to a supercomputer;
• Software: specialized modules that perform specific tasks;
• Mass Storage
• Image displays
• Hardcopy: laser printers, film cameras, heat-sensitive devices, inkjet units, digital units;
• Networking