In simple terms, Computer Vision refers to the field of study which deals with the process of giving the machines the ability to SEE & UNDERSTAND the content of digital images & videos, similar to humans.
It all started billions of years ago when a random mutation made organism, sensitive to light. Since then, it evolved from an accidental gene to an evolutionary gift. Being able to see helped all kinds of life to survive and explore vast distances. Throughout the years, the vision of humans and other animals has improved tremendously. With this, they also gain the ability to understand & respond to what they are seeing.
Although the vision was an evolutionary breakthrough, humans were successful in passing this ability to our gadgets.
Nowadays we have smartphones & cameras which have gotten even better than humans when it comes to seeing and capturing things.
But that's not enough... Just being able to see isn't much use, understanding what is currently being seen is what MATTERS MORE in Computer Vision.
This was something humans are unable to achieve for the past 40 years.
A normal human eye may look very complex but it works in a pretty similar fashion of capturing and absorbing light. Luckily, humans were able to reproduce this ability through the use of lens and paper film.
But after the light signals are registered by the eye, it is sent to the brain for further processing. Unfortunately, humans are still unaware of how the brain understands and process the visual information hence this ability became really difficult to reproduce through computers & mechanical components.
After decades of research and study, we have something somewhat useful although it's not perfect & far from being the solution humans need.
Computer Vision takes in some digital image like a photo or video and process & analyzes the position, colour, size and density of all the pixels on the image. It then tries to make sense of the pixel data by creating patterns through machine learning. The pixel information is stored in the form of a multidimensional array. This process is repeated numerous times to improve the accuracy of the computer vision model. In the next phase, our model will be able to compare a new image with previously trained data & recognize the content of the image.
However, this doesn't work very well all the time. In the real world, the model may encounter visual input which it may not recognize and hence there is a high chance of failure in such uncertain conditions.
Computer Vision even though is imperfect and in early stages, it has started to make its impact in various sectors helping people and nurturing their lives.
Here are some of the most impactful real-world applications of Computer Vision:
It's easy for your smartphone's camera to detect your face but it may or may not recognize without any additional hardware and software support. Things will change pretty soon as the facial recognition technology is improving every day and now machines are easily capable of recognizing you from a busy street through CCTV.
As AR technologies keep improving, humans will one day be able to fill in the gap between real and virtual. Mostly in, but not limited to games.
There are some places humans cannot go especially in pandemic & hazardous places. However, drones and other flying objects which do not require human pilot can go without the risk of being anyone hurt. It can be used to deliver supplies, medical kits. It can keep an eye on people through cameras, spray disinfectants in infected areas and so much more.
The future of automation is here. Everything you hate doing today will be automated in future someday. It will save a lot of time, money and effort. Most of these works will no longer need any human attention.
Self-driving cars will take on the highways and will help prevent a lot of accidents, transport passengers without any need of a driver, reduce traffic and increase efficiency.
OCR will improve and help us transcribe text which cannot be read normally. Our machines and gadgets will be getting better at reading.