What is the difference between image processing and computer vision? Both these disciplines pertain to images. And that is the only common denominator. Computer vision and image processing are two completely different tools used for various purposes. In this article, we will take a closer look at each one of them and see if machine learning can be any helpful.
Both image processing and computer vision are always about an image. However, here applies a comprehensive definition of an image. It can refer to photographs or videos captured by a camera, radar, ultrasound, or even a voice image recorded via a microphone. In general, we can think of three types of images:
· One-dimensional (a signal of amplitude over time, for instance, mono-channel sound waves)
· Two-dimensional (pictures and photos made up of rows and columns of pixels)
· Three-dimensional (videos)
Both image processing and computer vision can relate to any of these types of images. But what exactly is the difference between image processing and computer vision? Time to find out!
Computer Vision vs. Image Processing
In image processing, an image is, as its name indicates, processed. It means that at least one transformation is applied to an input file. And this can be done by a human with the usage of the dedicated software (to name just Photoshop, InDesign, GIMP, Gravit, CorelDRAW and many more).
Some transformations are done automatically. For example, sharpening, contrasting, filtering, and edge detection. All of them happen 100% automatically. A graphic just has to start a given operation. Other transformations are done manually, and these can be resizing, stretching, enhancing, and adding new layers or texts. These processes require much more attention and activity from the graphic. In image processing, you start with an image X, process it, and, as a result, obtain image Y. What transformations are needed depends on the context, purpose, and issue to be solved?
When we talk about computer vision, it is a different story. In computer vision, an image or a video is taken as input, and nothing happens to the file itself. The goal is to interpret the image and its contents. Computer vision may indeed use some of the image processing algorithms to solve its tasks, but the processing is never a primary goal. Actually, the image processing methods are harnessed for achieving tasks of computer vision.
One of the most important applications of computer vision is in the motor industry. Computer vision is used here as an assistant for the driver, especially in poor weather conditions. It scans the car’s surroundings and analyzes for potential threats, obstacles, and other relevant situations that a driver needs to react to while driving, to name just a pedestrian crossing the road.
Do computer vision and image processing use machine learning?
The short answer is yes. However, it is much more advanced and developed in computer vision. Consider Google Lens. This is a remarkable example of machine learning in computer vision that anyone with an Android smartphone can check instantly. Google Lens is an app that uses some image processing techniques along with machine learning technologies to give you more information about the object you’re pointing at. Google Lens detects an object, interprets it, and provides you with the results. The only thing you have to do is to point your smartphone’s camera at a specific object and take a picture! Google Assistant will tell you more about the object you’re pointing at. And it goes much further! You can take a picture of a given object and instantly search for it in Google search engine, check its model name and price, or even automatically translate a text written in a foreign language into yours.
And what about image processing? In the last article, we talked about medical image analysis. It is a perfect example of machine learning in image processing. The healthcare sector uses image processing, mostly for cancer and other disorder identifications. Two main techniques in use are segmentation and texture analysis. The machine learning algorithm can, for instance, detach the portion of the image that shows cancer cells, enlarge it and enhance the quality. This is a significant improvement in a physician’s work! Similarly, machine learning algorithms scan medical images for marks and distortions and analyze them if any are found.
Computer vision in the motor industry
As we established, the motor industry is one of the leading fields, where computer vision finds application. Consider some examples. Did you know that over 3,000 people die every day in a traffic accident? Far too much, and computer vision is one of many tools to solve this problem. The computer vision technology can be used to deal with a distracted driving issue.
The U.S. Department of Transportation’s National Highway Traffic Safety Administration (NHTSA) estimates that more than 3,000 automobile-related fatalities result from driver distractions. And everyone who has ever driven a car after a bad night’s sleep can confirm that–it’s very dangerous! The computer vision technology can help you stay awake and identify when you’re too sleepy or weary to drive. The computer vision application can continuously monitor your condition, basing on your eye state or head movements. The image recognition and computer vision technologies could pinpoint the exact moment when you’re not focused on the road and falling asleep. Your car immediately sends you an alert (a signal or a vibration) to get you back on the right course or advise you to take a nap before continuing to drive.
Computer vision protects pedestrians
Another example. Take pedestrian accidents, which are a severe problem. An estimated 6,227 pedestrians were killed in 2018 in the United States alone! Computer vision-powered car cameras can detect a pedestrian approaching a crossing before the driver may notice them and give a real-time alert. It is a significant assistance in driver’s work! One of the companies developing such a system is Volvo, known for years for taking care of the safety of all road users. They have developed a system called CWAB, which stands for Collision Warning with Full Auto Brake.
This system reacts when a pedestrian walks out in front of a car. If the driver does not take any action, the CWAB system will instantly activate the car’s full braking power. This system originates from the previous stages were Volvo was trying to help the driver avoid collisions with other vehicles. The current “version” also focuses on a pedestrian’s safety and gives instant access to the car’s full braking power. In the previous CWABs, it was only 50% of the car’s braking power.
To sum up, computer vision is a true game-changer in the motor industry. It assists a driver and protects pedestrians. When this technology is common in every vehicle on the road, we can expect a considerable decrease in car accidents.
Image enhancement in medicine
Image enhancement is a technique used widely in modern medicine to improve image quality and perceptibility. Medical imaging uses this for reducing noise and sharpening details to enhance the visual representation of the image. This technique includes both objective and subjective enhancements. As it turns out, many medical imaging methods, such as CT, MRI, or X-ray, suffer from low contrast. That leads to deterioration of image quality. This is why image enhancing is indispensable.
In the book called “Soft Computing Based Medical Image Analysis” you can find a precise and exhausting explanation of the image enhancement process in medicine. We can read there as follows:
“The enhancement in medical images happens by transforming HSV Space (Hue, Saturation, and Value), and Adaptive Histogram Equalization. The input color image is converted from RGB to transform HSV space while enhancing only the S space with an enhancement factor. The S and V spaces are subjected to Adaptive Histogram Equalization with a calculation of local variance for both. Further, the correlation between V and S space is calculated with luminance enhancement saturation feedback. Finally, the Enhanced Luminance V and S Space with H space are converted back to RGB to obtain the enhanced image. The gray images are subjected to the same procedure using Adaptive Histogram Equalization along with pre- and post-processing filters while excluding the conversion space.”
Thanks to the image enhancement process, a physician can make a quicker and more accurate diagnosis. The output file is much more legible and explicit.
The future of image processing and computer vision
Talking about the future, we stay with the healthcare sector. The next step in medical image processing is transforming 2D images into 3D models. In general, 3D imaging is a process where a standard 2D picture is converted into a 3D image by creating the illusion of depth. Future image processing will comprise rendering of colors and textures into the 3D model to make it look more realistic than today. With such expressive 3D images, physicians will be able to examine extremely high-quality 3D models of organs and tissues. This, in turn, will noticeably help them to carry out delicate surgeries and make accurate diagnoses.
And what about computer vision? We can expect that future computer vision will be used in conjunction with other deep learning technologies and other subsets of artificial intelligence to build more potent and advanced applications. Computer vision will play a significant role in the development of artificial intelligence in general. For instance, currently, there are speculations that sooner or later, humanity will be able to create an AGI–Artificial General Intelligence. In brief–that’s AI that matches human intelligence. Of course, this is the thing of a distant future, but the development of computer vision is one of the indispensable components to make it possible.
As always, we encourage you to talk with us about implementing artificial intelligence to your company. If you are wondering if modern technologies such as deep learning, machine learning, business intelligence or computer vision might be helpful–let us dispel your doubts! Give us a call and let’s talk about your company’s needs!
 Neuromation. How Computer Vision Can Change the Automotive Industry. Aug 8, 2018. URL: https://medium.com/neuromation-blog/how-computer-vision-can-change-the-automotive-industry-b8ba0f1c08d1. Accessed Dec 11, 2019.
 Phil LeBeau. Pedestrian deaths hit 28-year high, and big vehicles and smartphones are to blame. Feb 28 2019. URL: https://www.cnbc.com/2019/02/28/pedestrian-deaths-hit-a-28-year-high-and-big-vehicles-and-smartphones-are-to-blame.html. Accessed Dec 11, 2019.
 Volvocars. Pedestrian Detection (Collision Warning with Full Auto Brake). Dec 23, 2008. URL: https://www.media.volvocars.com/global/en-gb/media/videos/18529/pedestrian-detection-collision-warning-with-full-auto-brake. Accessed Dec 11, 2019.
 Satish S.Bhairannawar, Soft Computing Based Medical Image Analysis, Academic Press, 2018.