By: Annie Hu
An important aspect of science is analysis, which is what scientists do in order to make sense of their data and to see where their findings fit into the context of what is already known. Being able to find patterns and to communicate those results is a huge part of science. How could a series of measurements be turned into useful information? Well, through the magic of data processing, you have the ability to turn the measurements into some sort of graph. This information, then, could be interpreted much more easily by other scientists and people. Analysis is the way by which numbers, facts, figures, and images are brought to life and given meaning. After all, it’s not just data collection that constitutes what science is: rather, data collection is only a part of a much bigger picture.
Speaking of pictures! Just like the old saying goes, “one picture is worth a thousand words,” images acquired from experimentation can be extremely important reservoirs of information. But how can raw images gathered from experimentation be analyzed? Beyond the human eye, what other information are we able to gather? Many raw images are not directly usable because objects might be unclear or blurry, or image quality may be low. This is where image processing comes in! As a result of advancing technologies, not only can we manipulate images to obtain easier-to-analyze data, but we can also extract useful information from them. Digital image processing software can help scientists carry out many important functions. For example, it can allow scientists to visualize certain information present in the image, making data analysis much easier and simpler. The ability of digital image processing and analysis to gather useful information from an array of pixels is immensely useful, since computers are way better at all the counting and well, computing, that go into extracting important information from images than humans.
There are three main steps to image processing: acquiring and importing the image, enhancing and analysing the image, and then outputting it. The image acquisition stage usually involves doing some preprocessing. This means conducting necessary first steps, such as scaling the image, converting it from color to grayscale, or the other way around. Preprocessing prepares the image for enhancement, which is where things get interesting. There are many different techniques and methods of enhancement, but the end goal is generally to highlight certain features that are of interest in the image. One way of doing this is by the use of filters. A filter sorts out the unwanted components from those that are useful. A lot of math and algorithms are involved behind how filters work to bring out the relevant parts of an image. An image is essentially a two dimensional array of pixels, a matrix with each pixel having a certain x and y value. Each of these pixels also has a pixel intensity value, which is what image filtering changes, allowing color/tone to change (not position). When filters are applied, they move along the matrix, applying whatever specified algorithm. Simply put, by the end, we have an image changed and enhanced in some way. Different filters exist for removing noise, sharpening edges of objects, smoothing the image, creating contrast, etc. Enhancement is what allows us to take virtually any image that wasn’t initially fit for use and to make it possible to analyze.
Many image processing softwares also have numerous helpful built-in tools. These include features that are able to collect actual statistical data about the image that researchers can use to assess the image. For example, scientists can use software to take the histogram of an image. A histogram is basically a graph of the pixel intensity values. Histograms plot the number of pixels that have a certain tonal intensity. With this graphical representation, scientists can see the general distribution of intensities very clearly, which can be extremely helpful for certain analyses. Lastly, outputting an image involves simply saving the image file to a suitable format, which image processing software tools make an easy process.
Image processing is used a lot in the medical field for medical imaging and detecting disorders and diseases. Certain techniques and analyses are applied to images obtained from examinations like MRI scans and X rays to obtain the most comprehensible images possible. Sometimes image processing can also be applied to ultrasound images to enhance detail and edges. Image processing is also very applicable in other very interesting fields, such as forensics, where image processing is used to help define features by using techniques to detect edges and identify patterns. This could help detectives identify suspects. Image processing techniques are also incorporated and utilized in robot vision, engineering, computer science, microscopy, astronomy, and so many other amazing sciences!
Image processing is such an exciting frontier in science, and I hope you enjoyed learning about image processing, the applications, and the techniques involved in it as much as I did!
Raw images - Images collected straight from experimentation, not yet processed
Visualize - Make something visible
Array - Data structure consisting of elements
Image acquisition - the retrieving of an image to process
Scaling - the process of resizing a digital image
Grayscale - Black and white
Algorithms - a process or set of rules for the computer to follow
Pixels - The individual elements in the image array are known as pixels. A point in the image.
Pixel intensity value - The numerical value for the amount of light present in the individual pixel (1 value for grayscale/3 for RGB)
Noise - Random variations in pixel brightness. This will cause a “grainy” texture to the image.
Image Format - Standardized means of saving and storing images. For example, .jpg and .png are image formats.
What Did You Learn?
1. Why is image processing important and what is it used for?
Images are a very important source of information and evidence from experimentation. Gathering useful information straight from raw images, however, can be very difficult. Hence, image processing is crucial for making raw images easily analyzable. Most raw images are not immediately suited for analysis, so image processing is utilized in order to visualize details that are fuzzy or even hidden in the original image.
There are many applications for image processing techniques in research. Image processing is used in medical research to bring out the detail in images taken from medical imaging. This can be used to better detect things like cancer and other disorders. Image processing is also important in forensics (to identify suspects), robot vision (detecting edges), and other fields like microscopy (to get better images and evidence), and astronomy (once again, to enhance detail in the images).
2. What are the steps to image processing?
The main steps are image acquisition and importing, enhancing and analysing, and then finally outputting the image. Image acquisition involves some preprocessing to prepare the image for later steps. Enhancing is the step where images are processed to bring out certain parts more and to make details clearer so that the image can be analysed (enhancement and analysis go hand in hand). Finally, to output the image means to save the image and export as a suitable image format.