To most people, a digital image is just a picture. But to a computer, an image is a tapestry of pixels to be broken down and analyzed. The extraction of information from still images and even video, using image processing techniques, including algorithms, will be monetized shortly.
There are two forms:– Machine Vision, 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 to help developers 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.
When humans look at an image, they can identify every object in it (left), whereas most computers see only what is most salient (right). Facebook AI Research aims to push machine vision toward a similar goal: understanding images and objects at the pixel level.
Computer vision uses deep learning models, such as convolutional neural networks (CNNs), to analyze images and video streams, enabling applications such as object detection, facial recognition, and automated visual inspection.
Here’s a quick look at computer vision data analytics and its current uses:
- Computer Vision’s primary role is to determine whether an image contains a particular object. Algorithms automatically analyze images and extract information.
