Himani Sharma elaborates upon the concept of computer vision or the machine vision, which allows to undertake an analysis of a three-dimensional object, by mechanically detecting the patterns therein to make inferences.
Computer Vision, also termed as machine vision, is a processing of images from the real world with the help of computers. Basically, it recognizes the structure or various patterns that are found inside an image for further analysis. One can understand that by using video cameras, from which one can get the information of the surrounding environment and process it. It is used to produce a structural analysis of a 3D view.
Computer Vision Vs Human Vision : Normal human sight is limited to visible spectrum, memory is very less compared to that of system etc.
Two techniques are widely used for used for computer vision:
- Image processing (transformation of one image into another image)
- Pattern Recognition (use of various mathematical techniques to produce different patterns from the input data)
Fig. Computer Vision
SPECIFIC TASKS INVOLVED IN THIS TECHNOLOGY
1. Recognition : To illustrate whether or not the image contains some specific information. Recognition problems include – object recognition (google glasses provide the solution), identification (a particular instance of an object like fingerprints).
2. Motion Analysis : To produce an estimation of the velocity of the image at points or of the camera, by either -Egomotion (rotation and translation of camera) and Optical flow (movement of points in an image).
3. Scene Reconstruction : It is done when one or more images are present to produce a 3D model of the scenario.
4. Image restoration : Basically includes the usage of filters to separate the noises present around from the image to be processed.
APPLICATIONS OF COMPUTER VISION
1. Automotive driver assistance and traffic management : Traffic management can be done with the help of the news number plates that are developed recently in India. Image processing can be done using roadside cameras etc. Brand like BMW, Volvo etc. in Israel use mobile eye for active braking to assist the driver. System is developed to track drowsiness factor for drivers and alarm the people inside the vehicle about the major cause of road accidents these days.
2. Medical and Biomedical : Visually impaired people can benefit from special glasses, mobile devices to assist them. Moreover detection of surgical tumors is carried out in many areas. Techniques are being developed to convert the gestures or hand movements, which are done for deaf and dumb people, into voice signals using computer vision.
3. Industrial Automation : For food and agriculture industries to monitor bake color, size, type of bread and cookies etc. and for textile and printing industries to detect and monitor the quality of the printing and material produced. Bar code reading and labels of packaging reading (iphone scanner) are also used.
4. Gesture Recognition : While playing games, human motion and gestures are recorded for control of computers and other connected devices.
5. Film and Video : For match updates, tracking of the players and the ball is done in tennis, cricket etc. using high speed cameras. For the film industry facial expressions of the audience is recorded for the main lead characters of the movies and thus popularity is detected.
6. Object Recognition and AR for Mobile Devices : For mobiles cameras, Amazon has developed a system for the search of the product, based on the image provided to them. Wikitude company of Austria provided the feature of ‘augmented reality’ for mobile phones.
7. Security and Biometrics : Tracking of people within the malls, stores and markets for security and also marketing and services. A France – based company, The Poseidon System, monitors swimming pools to warn about accidents and drowning victims. Biometrics is used for facial recognition.
8. Military : The application is in the form of detecting the missiles, vehicles of soldiers and movement of enemies. Moreover, during a war or a scenario on battlefield, image sensors help or guide about the conditions so as to take strategic conditions.
9. Submersibles : They are a form of autonomous vehicles like robots with wheels etc. for providing support to the pilot and driver. They are heavily used in missile guidance and other space exploration fields.
Application : Mobile phones are already using image processing to some extent, and for educational sectors virtual teaching classes can be there.
Moreover, an HR robot can be there to take an interview and sense the actions and expressions of the interviewee.
FUTURE SCOPE
We all have seen that the computers are able to detect almost every face it gets across. The only problems are the pose or the structure of the person, which may be in an inverse stance or he/she is surrounded by noise. But this will surely be eliminated in coming time.
In its development, funding is provided by various renowned companies like Microsoft, Intel, NEC and Kodak. Many fields like the social media, where huge number of photos are uploaded and videos (on YouTube) which need to be managed, are pushing the growth of computer vision. On common grounds. Internet search engines, industrial inspections and other medical systems are using this technology. As we all know that startup companies are increasing these days and they are also motivating the use of computer vision as a solution to many image related problems.
