According to Gartner,Inc 2019 CIO Survey, the percentage of Enterprises Employing AI Grew 270 Percent Over the Past Four Years. Why this happened? The major reason behind this is the big jump of AI capabilities. General AI is still far from being achieved, but AI itself has matured significantly so more business is more ready to accept and adopt AI technologies.
The military and government have always been collecting numerous open-source images and videos and need technology to help detect forgeries. For example, whether a photo is authentic or photoshopped is difficult to identify just by our naked eye, then AI has a role to play. That is so meaningful because this world is swarming with so called ‘fake news’, so the intelligence community definitely needs to filter information to make the right decision and consult the right person such as professor Edward Delp.
According to IEEE Xplore, Edward Delp is a professor of Electrical and Computer Engineering and also a director of Purdue’s Video and Image Processing Laboratory, or VIPER. He has consulted for various companies and government agencies in the areas of signal, image, and video processing, pattern recognition, and secure communications.
According to information posted by Emil Venere, researchers from seven universities developed an “end-to-end” system which allows handling massive information uploaded regularly to the Internet, and this project is funded by DARPA with $4.9 million. And the initial conclusion is that machine learning can indeed detect which part in the photo has been modified.
This research is not only essential for media forensic, but also will bring about change in wider academic fields, because scientific research papers used to use fake photos to back up their fabricated experiment data.
Illegal logging and farming
Illegal logging and illegal trade have always been headaches for law enforcement. However, now forests can be better protected by AI computer vision technology. Drones can be deployed to monitor, track and gather logging activities by capturing images. Then AI will generate a unique fingerprint to track and trace every single tree.
For example, some computer vision scientists right now are trying to create unique identities for every timber so every timber harvested can be traced from the beginning to end customers throughout the supply chain.
When the ring pattern on the timber cut matches the pattern on the timber stump stored in the cloud, the timber logged can be authenticated and be traced to its logging point. This technology is not only useful for preventing illegal logging, but also helpful for farming industry and food safety. Dimensions for identity will vary for different agricultural products for sure.
Microsoft Research Cambridge has been developing a project called InnerEye which harnessed AI technology for automatic, quantitative analysis of three-dimensional medical images. It claims can visually identify and display possible tumors and other anomalies in X-ray images. It usually takes one or more hours for a specialist technician to examine and determine the results of CT image scans, but InnerEye technology can reduce that to several minutes.
Below is a video from Microsoft Research on Youtube about how it works:
This project employs algorithms such as the latest Convolutional Neural Networks for the automatic, voxel-wise segmentation of medical images. The mission of this project is to democratize medical imaging AI. Therefore, Microsoft has made InnerEye Deep Learning Toolkit open-source for many individuals and organizations to access.
In fact, AI backed computer vision can help the health care sector with more tasks such as tracking actions in an OR, facial authentication of patients and so on.
AI vision can be applied for different purposes in the sports industry from player pose tracking, motion capture, stroke recognition, behavior analysis, coaching to automated media coverage.
According to some sports performance analysis, NFL in United States has already adopted computer vision to help analyze offensive formation during a play. During a game, coordinates of players, every run or pass of a player will be tracked to help coaches to analyze oppositions’ tactics. In the 2011/2012 football season in Germany, an automatic player tracking system with two cameras has been used in the press area of any stadium, which reduced the number of operators required to get accurate and latest data.
Color-based segmentation algorithms are currently being used to help cameras to distinguish the grass from, for example, a player wearing green sportswear. With the latest algorithm, tracking system is good at treating grass as the background and detecting the motion of the segmented foreground players.
With the right computer vision technology, a level playing field for sports industry can be better ensured.
——————–the end ——————————–
Who are we?
Image Cut Out | API | Remove Background
PicUP.AI is an artificial intelligence company which is focusing on automatic image and video processing. We offer API solutions for companies with batch photo editing needs such as image cut out, removing image background, photo retouch, cartoonizing, etc.
For more information on our product and services, please click here.