Ever since its inception in the early 1960s, Computer Vision has been one of the formidable forms of pattern recognition that computers are still yet to master. Essentially a digital image detection process, Computer Vision enables machines to identify images based on pre-existing data and use it to classify new images just like how the human eye and visual cortex of the brain coordinate to contextualise images by interpreting both abstract and visual meaning of images.
Commonly abbreviated as CV, Computer Vision is a branch of AI that improvises itself through algorithms based on neural networks. That’s why developing computer vision requires a large amount of input data and pattern-recognition assistance so that images can be distinguished according to their specifications. The use of CV also extends to facial and fingerprint recognition, object tracking, surveillance, and is widely used in healthcare technology such as X-ray and CT scans.
Social networking sites like Instagram, Facebook, Snapchat and video-sharing platforms like YouTube, Vimeo, TikTok and Dailymotion use Computer Vision to aggregate, classify, recognize, and label images that are uploaded, shared, and hosted by users from all over the world. Instagram, world’s most popular photo-sharing app owned by Facebook Inc., is reported to have nearly 95 million photo uploads a day while YouTube has 500 plus hours of content uploaded on its site every single minute.
To process such large amounts of visual data on social media as well as e-commerce sites, computers need to “see” and “read” digital content by understanding what the object is, how it generally looks like and what it means. Computers define these characteristics by extracting pixel and text related data from the images as well as deciphering the physical dimension of the image.
When computers process an image, they first standardize photometric properties like brightness, light source, camera position and then focus on image-centering. Optimisation is the next stage in Computer Vision and is carried out by removing digital artifacts that minimize the representational quality of images.
Prior to the adoption of Deep Learning, computer scientists had to manually build programmes to perform Computer Vision tasks. This involved actual measurement of facial features and their distance from each other, labelling and feeding in examples of other faces or images so that computers know what to look for.
Today, Machine Learning and Deep Learning automate searches through layered neural networks, producing real-time results from a data source of billions of images. However, Computer Vision still has a long way to go when it comes to recognising abstract meanings from images. Images seen by the human eye are associated with feelings of warmth or apathy – a function that is determined by the human brain. Computer Vision merely cross checks batches of data with a ready reckoner, unable to interpret “mood” or “emotion-related” concepts in the image.
Computer Vision has a multitude of target applications in daily life. From Optical Character Recognition in scans to CGI in movies, CV plays a dominant role in military intelligence, biometrics and even assorting images on Google Photos. It’s also one of the building blocks for embedding augmented reality in devices like smart glasses whose global market is set to reach 31. 1 million units by 2027.
Applications of the technology are widely used in sports, specifically in cricket in India, where strokes, dismissals and ball trajectory are tracked to high precision. In healthcare, Canadian startup DarwinAI along with researchers from the University of Waterloo recently released AI software capable of detecting Covid-19 through chest radiology imagery. CV can also measure cancer progression, tumour sizes and aids diagnosis and recovery.
With visual data market exploding because of bigger computing storage and extensive use of visually assisted devices, it has wide expansion scope in several industries. In 2019, the global computer vision market size was valued at $10.6 billion – a figure that is expected to soar at a CAGR of 7.6% from 2020 to 2027.
Research advisory firm AIMResearch in its July 2020 study titled ‘State of Computer Vision in India – 2020’ pegs the Indian Computer Vision market at $2.46 billion. According to the report, MNC Tech category grossed a 47.2% market share valuing $1159 million, the highest among represented sectors. Domestic IT companies followed with a 30.5% market share, valued at $749.6 million.
The report also maps the CV job market which has nearly 22,500 open positions across enterprises in India. The Indian Institute of Information Technology Sri City, Chittoor (IIITS) established by the Ministry of Human Resources Department is one among the 20 institutes designated by the government to focus on IT, education, research, and development. Their Computer Vision Group enables research on applied and theorical CV, often undertaking projects funded by the government.
Ever since its inception in the early 1960s, Computer Vision has been one of the formidable forms of pattern recognition that computers are still yet to master. Essentially a digital image detection process, Computer Vision enables machines to identify images based on pre-existing data and use it to classify new images just like how the human eye and visual cortex of the brain coordinate to contextualise images by interpreting both abstract and visual meaning of images.
Commonly abbreviated as CV, Computer Vision is a branch of AI that improvises itself through algorithms based on neural networks. That’s why developing computer vision requires a large amount of input data and pattern-recognition assistance so that images can be distinguished according to their specifications. The use of CV also extends to facial and fingerprint recognition, object tracking, surveillance, and is widely used in healthcare technology such as X-ray and CT scans.
Social networking sites like Instagram, Facebook, Snapchat and video-sharing platforms like YouTube, Vimeo, TikTok and Dailymotion use Computer Vision to aggregate, classify, recognize, and label images that are uploaded, shared, and hosted by users from all over the world. Instagram, world’s most popular photo-sharing app owned by Facebook Inc., is reported to have nearly 95 million photo uploads a day while YouTube has 500 plus hours of content uploaded on its site every single minute.
To process such large amounts of visual data on social media as well as e-commerce sites, computers need to “see” and “read” digital content by understanding what the object is, how it generally looks like and what it means. Computers define these characteristics by extracting pixel and text related data from the images as well as deciphering the physical dimension of the image.
When computers process an image, they first standardize photometric properties like brightness, light source, camera position and then focus on image-centering. Optimisation is the next stage in Computer Vision and is carried out by removing digital artifacts that minimize the representational quality of images.
Prior to the adoption of Deep Learning, computer scientists had to manually build programmes to perform Computer Vision tasks. This involved actual measurement of facial features and their distance from each other, labelling and feeding in examples of other faces or images so that computers know what to look for.
Today, Machine Learning and Deep Learning automate searches through layered neural networks, producing real-time results from a data source of billions of images. However, Computer Vision still has a long way to go when it comes to recognising abstract meanings from images. Images seen by the human eye are associated with feelings of warmth or apathy – a function that is determined by the human brain. Computer Vision merely cross checks batches of data with a ready reckoner, unable to interpret “mood” or “emotion-related” concepts in the image.
Computer Vision has a multitude of target applications in daily life. From Optical Character Recognition in scans to CGI in movies, CV plays a dominant role in military intelligence, biometrics and even assorting images on Google Photos. It’s also one of the building blocks for embedding augmented reality in devices like smart glasses whose global market is set to reach 31. 1 million units by 2027.
Applications of the technology are widely used in sports, specifically in cricket in India, where strokes, dismissals and ball trajectory are tracked to high precision. In healthcare, Canadian startup DarwinAI along with researchers from the University of Waterloo recently released AI software capable of detecting Covid-19 through chest radiology imagery. CV can also measure cancer progression, tumour sizes and aids diagnosis and recovery.
With visual data market exploding because of bigger computing storage and extensive use of visually assisted devices, it has wide expansion scope in several industries. In 2019, the global computer vision market size was valued at $10.6 billion – a figure that is expected to soar at a CAGR of 7.6% from 2020 to 2027.
Research advisory firm AIMResearch in its July 2020 study titled ‘State of Computer Vision in India – 2020’ pegs the Indian Computer Vision market at $2.46 billion. According to the report, MNC Tech category grossed a 47.2% market share valuing $1159 million, the highest among represented sectors. Domestic IT companies followed with a 30.5% market share, valued at $749.6 million.
The report also maps the CV job market which has nearly 22,500 open positions across enterprises in India. The Indian Institute of Information Technology Sri City, Chittoor (IIITS) established by the Ministry of Human Resources Department is one among the 20 institutes designated by the government to focus on IT, education, research, and development. Their Computer Vision Group enables research on applied and theoretical CV, often undertaking projects funded by the government.
