To most they consist of just pixels but digital images, like any other form of content, can be mined by computers for data, which can then be analyzed. The extraction of information from still images and even video, using image processing techniques including algorithms, is on the cusp of going commercial.
There are two forms – Machine Vision, which is the more “traditional” form of this tech, and a digital world offshoot, Computer Vision (CV). While the first is largely for industrial use, cameras on a conveyor belt in an industrial plant being an example, the second is to teach computers to extract and understand data “hidden” within digital images and videos.
This August, Facebook said it was open-sourcing its efforts for developers to further develop its Computer Vision tech. The post by FB research scientist Piotr Dollar posted this image to explain the difference between human and computer vision.
Here’s a quick look at this emerging tech and its current uses:
- Computer Vision or Machine Vision’s main role is to determine whether or not data in an image consists of some specific object or activity. Algorithms are used to automatically analyze images and extract information
- While a lot work’s been done to apply computer vision techniques to extract info from video, automated computer vision is not yet reliable enough to detect anomalies or track objects. Hence, in such cases, humans are still in the analysis loop to assess the situation
- Certain global companies including e-commerce sites have already started using image analysis in their predictive analytics efforts to forecast what their customers will want next
- One field where Computer Vision is being used is for improving the results of Internet search engines, since this tech can contextualize images in searches. Which means it can tag and identify objects, and even the settings in your uploaded photographs
- Facial recognition techniques used by social networks and other companies for tagging people in photographs is based on CV. Facebook has taken this tech to such a level that it can, with its facial recognition software, identify a person even if his/her face was covered up
- The applications of CV include augmented reality, biometrics, face recognition, gesture analytics and robotics
- There are developers about to release search engines for image sharing services such as Instagram that will sift through and make sense of the data from the stream of photos uploaded by millions of users
- CV is used extensively in remote-controlled drones
- Some IT companies like Microsoft, for example, which has a Cognitive Services division, offer APIs to help anybody extract rich information from images in order to categorize and process visual data
- A few retailers in the UK and the US have started using Computer Vision to help potential customers to virtually “try out” furniture and wallpapers within their own room before placing an order.
Have something more to say on this blog post or topic? Write in to us firstname.lastname@example.org