3 Applications for AI Image Recognition

September 23, 2017 by Heather Hamilton-Post

AI image recognition has uses across industries, from the personal to the professional.

AI image recognition has uses across industries, from the personal to the professional.

Artificial intelligence is constantly changing the way we view the world--or in some cases, the way we no longer have to view it. Where humans once had to pore through photographs and video footage to pick out faces in a real life search for Waldo, machines are now more capable than ever of learning to identify them and pick them out from a sea of other faces across a number of visual mediums.

Image Organization Across Platforms

Few people think of AI when they consider things like personal organization, but it has the potential to change the way video and images are organized. Photo organization apps are using AI to better store and separate photos, offering search and discovery functions that drastically improve usability. These apps use image recognition APIs to offer image tags, auto-generated keywords, and automatic categorization based on visual topics, often across devices.

On websites for stock photography, AI image recognition better allows users to upload and appropriately tag content, which allows users to actually find what they’re searching for.

Image courtesy of Pixabay.

Law Enforcement and Intelligence Agencies

BrainChip, a developer of both software- and hardware-accelerated solutions for advanced artificial intelligence and machine learning, recently announced the release of the BrainChip Accelerator, a hardware acceleration board. The board is an 8-lane PCI-Express add-in card designed to increase the speed and accuracy of the object recognition function of BrainChip Studio software and up the available video channels to 16 per card. Brainchip prides itself on offering a low-power card with easy installation in video surveillance systems without the need to upgrade power systems or thermal management.

BrainChip Studio is known for assisting law enforcement and intelligence organizations in picking objects out of video streams, both archived and streaming. The card allows organizations to search multiple video streams more quickly, speeding up the overall process. The system learns by looking at one low-resolution image, even if it is low-light, etc.

Six BrainChip Accelerator cores in a Xilinx Kintex Ultrascale FPGA take care of processing, and each core performs user-defined image scaling, spike generation, and spiking neural network comparison for fast (and more probable) object recognition. BrainChip’s Senior Vice President of marketing and business development Robert Beachler estimates that there are four exabytes of video data stored in video surveillance systems. “In surveillance, speed and accuracy of analysis are critical concerns for law enforcement and security agencies. The ability of BrainChip Accelerator to process video frames six times faster, while improving the accuracy of object recognition, is a significant force multiplier. It is also a further demonstration of the valuable role that artificial intelligence can now play in these applications,” he said.

BrainChip Accelerator works with both Windows and Linux and is currently available to law enforcement agencies as an integrated server appliance or as an add-in card for security integrators and OEMs.

Image courtesy of BrainChip Holdings Ltd.

Social Media

AdWeek reports that approximately 3.2 billion photographs are shared every day across social media as the push for visual content is stronger than ever. Brands are searching social media for images related to their brands, and social media marketers are emphasizing the necessity of people in said images, which are more likely to generate engagement and be shared by others. In this quest to create viral content, brands are able to capitalize on AI without relying on users to tag or use appropriate keywords.

Salesforce is getting in on the action too--Rob Begg, Vice President of marketing for social and advertising products at Salesforce is excited by AI’s newfound prevalence. “If you think of it from a company point of view, there is a huge volume of tweets and [social] posts. What AI does best is help surface and source the ones that are relevant,” he told Tech Crunch. He explained that, while there may be thousands of posts about cars, only a handful are relevant, and AI helps find those. For Salesforce, Beggs believes it could provide insights into how products are being used, provide a way to track brand displays online hidden within pictures, and allow you to find out when influencers are using your products.

Image courtesy of Facebook.