Distributed AI With Data Privacy
Artificial Intelligence built on world leading research
VSL – The Next Generation AI Platform
Vision Semantics Limited (VSL) is a leading Deep Learning AI solution provider focusing on computer vision machine learning. VSL has developed proprietary video analysis and dynamic scene understanding software that enables dynamic search of the virtual environment to impact outcomes in the physical world.
In solving the most complex Deep Learning AI computer vision problems, VSL has developed a unique approach to Decentralised Machine Learning, the next generation AI platform, based on a dynamic lifelong learning AI model to “search and learn”.
Today, AI is based on big centralized data “learn and forget” algorithm models with limited data privacy, large computing processing power, and big power consumption. The edge is the next stage of the evolution of AI technology because of the physical constraints, the cost constraints, and the practical constraints of running all AI applications in the cloud. VSL’s approach, perfected in computer vision, enables it to deliver a decentralized AI platform, more advanced than Google Federated Learning.
It’s the future platform for AI with multiple sector applications.
Artificial Intelligence Products Built on
VSL began life as a spin out from Queen Mary University of London Computer Vision Department, led by founder Professor Sean Gong. The team has been conducting world-class research in computer vision and machine learning for over 20 years, internationally renowned for its work on unusual behaviour recognition, person re-identification, multi-camera tracking, and face analysis in video and images. This research has led to over 400 published papers, over 27,000 citations and filing of 9 families of international patents, with 34 underlying patents, providing barriers to competitors exploiting computer vision commercially.
The main areas of application include person re-identification, action and behaviour analysis, attribute recognition, clothing analysis, image and video super-resolution, crowd analysis and counting, video semantic search and annotation, video summarisation, and privacy by design image analysis. VSL takes this research and commercializes it so that it is robust, scalable and can be implemented by partners in real world situations.
Deep learning neural networks which transform the outcomes of computer vision analysis
VSL has developed world-class video analysis and dynamic scene understanding software that employs innovative computer vision and machine learning methods focused on deep learning. It has developed a commercial grade software platform for patented automatic video analysis of vast quantities of video & images running up to 100X real-time.
In particular, VSL has developed Person Re-identification (or person RE-ID for short) the ability to find rapidly a person in time and space in a vast quantity of video data from distributed cameras, even if there is no clear view of the face or lighting is dark. VSL is unique in being able to find people within low resolution and poor quality video footage.
AI Powered Person Re-Identification (RE-ID)
The video surveillance industry is experiencing a massive growth spurt as we increasingly use cameras to prevent crime, secure evidence, protect our property, and improve public safety in the home and at large. However, data generated from digital surveillance cameras has become so massive that it is often impossible or commercially impractical for human operators to make sense out of it in a timely manner.
A major challenge is Person Re-Identification (RE-ID). That is the process to find matching images of a person (in a gallery of images) for a given probe image whilst not using facial imagery features nor any other personal specific biometrics such as gait (which are often obscured). RE-ID explores features found on the entire body from clothing, style and carried objects. RE-ID is designed to work when there is no training example (Zero-Shot Learning). Face recognition works by exhaustive enrollment of training samples from every individual in a database (N-Shot Learning). Face recognition works only when the subject is close enough, well-lit and facing towards the camera. Usually in CCTV footage this is not the case, as a face is often looking away or covered from the camera. RE-ID does not use facial imagery and is inherently privacy preserving. It does not identify who is.
How Does Re-ID Work?
VSL solutions provide the ability to search for people when
facial recognition is not feasible within any unstructured and
uncontrolled environments, indoors and outdoors, close-by and
far-away, day and night. VSL can rapidly find people in time,
with accuracy over large-scale crowded environments that public
safety agencies operate in, with the ability to handle low resolution
RE-ID applies a Zero-Shot learning approach, a process by
which a machine learns how to recognize objects in an image
without any labelled training data to help in the classification.
In other words, Zero-Shot learning aims to help machines
categorize objects or interactions that they have never seen before. This makes the solution scalable - it can deployed in cities internationally without data training.
VSL uniquely applies RE-ID to human-in-the-loop “search and learn” AI algorithms, bridging transfer learning, reinforcement learning, semi-supervised and unsupervised deep learning, to enable rapid and scalable RE-ID in unknown target domains from an open-set world.
VSL RE-ID software has grown beyond research lab PhD experimental code to become commercially mature and tested. VSL RE-ID technology is underpinned by world-class RE-ID algorithms that have led the academic RE-ID benchmarking not only on supervised, but also on unsupervised and domain transfer testing, on both the largest benchmark Market-1501 and the hardest benchmark GRID. VSL has gone beyond market testing into making the technology scale for real commercial deployments.
VSL is the driving force in this market with a 12-18-month sustainable technology lead in RE-ID over the nearest competitor. The technology is backed by 9 international patent families with 34 underlying patents. This is combined with 10 years of commercial work with international Government and law enforcement agencies to ensure the technology is accurate, fast and scalable for real world deployments.
Capabilities and Roadmap
Today, RE-ID is applied to post event analysis where we search for a person in time and space over a network of CCTV cameras video footage. This capability is being enhanced by adding attribute search to RE-ID, so we can now search for a person with an object (bag, umbrella, etc.) and a vehicle (car, motor bike or bike).
Into this environment, VSL released Real Time RE-ID in 2020 and this transforms the capability of public safety to act in real time to trigger events picked up in video footage. If police and public safety officials can monitor and respond in real time, then a crime or social disturbance can be recognised and prevented.
VSL is currently working on a behaviour video analysis module which will work within the RE-ID framework. This will enable us to re-identify behaviours which serve as a trigger event to a predicted outcome. The module facilitates abnormal behavior profiling and can then facilitate event alerting to identification of lost or self-harm passengers and group social distancing violations. This will enable future automation of the process flow in protecting threats to the public.