Processing & Analysis of Digital Images
Digital image processing is the method and technology that process images to remove …
Digital image processing is the method and technology that process images to remove the noises, ima ge enhancement, image restoration, image segmentation, feature extraction and so on. In recent years, with the development of the technology of image processing and improving the more and more algorithms, the science of digital image processing has been the important study in some fields of engineering, computer science, information science, physics, medicine, etc.
There are two types of images broadly: vector and bitmap. The vector image (by mathematical vectors) allows resizing image at any scale without loss of quality, however, at the moment that must be printed for example, it is converted in a bitmap suitable to the characteristics of the equipment. The process of conversion from vector to bitmap is called rasterization. The bitmap image is – as the name suggests – mapped bits, where the image is organized into a series of rows and columns formed by pixels.
So the bitmap is the final type of the entire image displayed. Therefore, it is necessary to understand the meaning of resolution. Resolution in this case is the ability of a capture or reproduction images system has to reproduce details. The resolution is given in dots per square inch (dpi), so, how higher the resolution is, better the details reproduced by a given system.
Some common image file types are as follows:
- JPEG (Joint Photographic Experts Group): The JPEG file type is a file type that allows the use of various compression techniques, whereas most of them, present losses. In this case the original picture will not be exactly identical to the recovered image after decompression.
- PNG (Portable Network Graphics): The PNG file type emerged as an evolution of the GIF file type, supporting millions of colors, transparency and animation. This file type can be used in place of JPEG in some situations because, despite having compression, this does not generate losses.
Basic Guidelines for Best Practices In Image Processing
The increasing use of PADI led to universities and research centers around the world publish some basic and relevant information.
Here are some topics that should be used like guidelines for best practices in image processing DICOM- Digital Imaging and Communications in Medicine, ISO 5052, 523.
- Treating Images as Data: An image can be translated as a multidimensional array of values. Assuming that a bitmap image is as a real picture with two-dimensional and, the three- dimensional effects are achieved through painting techniques, so each pixel has its place in the matrix (such as a spreadsheet) and, in each location is stored one or more values.
- Save the Original: It is very important (and safe) that manipulated images can be compared with the original image (the first captured image) to make sure that important or relevant data to research does have not been lost or, which have not been entered new data (artifacts) in the processed images. Therefore, the original image must always be maintained.
- Making Simple Adjustments: As photographic images, simple adjustments are generally acceptable to help visualize the information that is present in the image, such as brightness, contrast and gamma (in moderation).
- Comparing Images: A precise comparison between the treated and control images will be reliable only if both are acquired under identical conditions. In this case you need to be especially careful about the configuration of image acquisition devices (electron microscopes, for example).
- Filters Degrade Data: The commercial software tools (well known) for manipulating images were created mainly for printing and design, and not for scientific research. The filters are mathematical functions called kernels of convolution.
Although there is already a real concern about the minimal care needed regarding image manipulation oriented to PADI – processing and analysis of digital images – very little has been done. The best practices presented are the intiial path: replace images of type JPEG with those of type PNG wherever possible, be aware of what makes a filter before applying to an image, and keep the original image safely. By following these basic guidelines, you are ready to start delving into the world of processing and analysing digital images. so that basic information (and practices) comes to the researcher who only uses PADI as a resource to power your own research. The best practices presented are the initial path: replace where possible images of type JPEG by PNG, be aware of what makes a filter before applying it to the image, keeping the original image safely.
By: Saleha Shujaat