The Eye of Intel: AI in Computer Vision
Mesum Raza Hemani
Leader in AI & Data Science (Karachi, Pakistan) - Founder Karachi AI
You are being spied! Yes.. The intelligent eye of the electric powered board is now more powerful than ever. Better in vision, better at decisions and better at speed. Today Computer Vision has transformed the way we interact with our devices with our faces being our identity, our cars drive smart without human intervention with a safer strategy, even we have new digital birds named as Drones that can fly with independent vision and task to accomplish.
Before your jaws start dropping in wonder.. lets discuss!
What is Computer Vision? : Computer vision is an interdisciplinary field that deals with how computers can be made to gain high-level understanding from digital images or videos. From the perspective of engineering, it seeks to automate tasks that the human visual system can do. Computer vision is concerned with the automatic extraction, analysis and understanding of useful information from a single image or a sequence of images
That was Roman! You might be thinking.. How does a computer see? Does it have a conscious?
So, let’s understand how do machines see?
A. Represent Color numbers: In computer science, each color is represented by a specified HEX value. That is how machines are programmed to understand what colors the image pixels are made up. Whereas as humans we have an inherited knowledge to differ between the shades.
Then..
B. Image Segmentation: Computers are made to identify similar group of colors and then segment the image i.e. distinguish the foreground from background. The technique of color gradient is used to find edges of different objects.
Then..
C. Finding corners: After segmentation, images are then looked up for certain features, known as corners. In simple words, algorithms search for lines that meet at an angle and cover a specific part of the image with one color shade. Features, called corners are the building blocks which help to find more detailed information in the image.
Then..
D. Find textures: Another important aspect to identify any image correctly is to determine the texture in the image. The difference in textures between two objects makes it easier for a machine to correctly categorize an object.
Then..
E. Make a guess: After implementing the above steps, a machine needs to make a nearly-right guess and match the image with those present in the database.
and Then..
F. Finally, see the bigger picture! At last, a machine sees the bigger and clear picture and checks if it was right identifying the one, as per the feeded algorithmic instructions. The accuracy has improved a lot in past years but still, machines make mistakes when asked to handle images with mixed objects.
before we certify you might love to learn this:
AND..
NOW THAT YOU HAVE CLEARED A CRASH COURSE INTO BASIC CV!
So I have heard this started back in some 60's, why is it important now? Whats new now that has led to the significant improvements in vision and recognition that has never been before..
Traditional Techniques for computer vision have been around since the topic started, with basic image segmentation algorithms to more improving and sophisticated task oriented algorithms that did best in what they were made for. But these techniques limited the scope of vision to be supervised only. Supervised means that Algorithm is either designed to tailor the conditions and specificity of the problem and every new case will require to be accommodated to be recognized.
In late 80's more sophisticated and generalized techniques emerged based on Neural Networks that required immense computing power to resolve vision problem in more general and yet accurate way. Since inception of Moore's law Computers have been able to multiple their processing powers more than ever being.. which on the parallel enabled heavy and complex machine learning algorithms to run. Hence enabling modern techniques for CV such as Convolution Neural Networks, Recurrent Neural Networks, Generative Networks and Many more.. and yet to come.
Today Computer Vision based on modern machine learning algorithms have been able to detect Cancers, help radiologist identify minor infections, helping navigate customers through store departments. Which include Face recognition?,Gesture recognition?,Image search?,Machine vision?,Optical character recognition,Remote sensing,Robots and Self-driving cars?
Now you have understood CV and its utilization and would be eagered to learn about commercial products that are built via computer vision, Why not? Lets Get Started ;-)
- Commerce:
2. Health Care:
One of the Famous personality in AI democratization tweeted..
3. Security/Surveillance:
4. Generative Arts:
And many many more.. one of the interesting startups we met was ConnectHear, that utilizes computer vision pose estimation and gesture detection techniques to interpret deaf commands spoken by a large community of speaking impaired persons.
Founded by three young bright Pakistani students, who have dreamed from the small to make a social impact using Technology.
ConnectHear is using AI to improve Deaf user experience and more accurate translations so that special emotions and daily commands for the community are connected to existing humans. You shall be curious to learn more about how they use, we would then recommend you to listen to our third Event: Karachi.AI Meetup # 3: AI in Computer Vision, IoT & Robotics.
I hope you must have been disrupted by this little insight over future and thinking furiously to learn about AI and how to dig into it: Well Karachi.AI is the stop for you :)
Welcome to Karachi.AI
Like Page : www.facebook.com/Karachidotai
Join Group : www.facebook.com/groups/karachidotai
Immerse with Complete Event: https://www.youtube.com/watch?v=9n1gG61_WAc
Presentations presented at the Karachi.AI 2.0 Meetup: https://drive.google.com/drive/folders/171xY4DK2cLO8bWWH62t60n9j4r_ARYmP
Authored By:
Mesum Raza Hemani
About: He is a chartered accountant who turned his back on practice and diverted his career along the lines of business analytics, data mining. Specializes in advance machine learning & deep learning for business via University of Washington and Stanford distance program. Founder of Karachi.AI & Lucid Data Artist
Thank you for Reading this article & if you like it share it with you acquaintances and friends.