Giving Your Business Sight
VisionX has filed for patent protection at the United States Patent and Trademark Office (USPTO) for our innovative camera intelligence technology that uses machine learning to classify human behavior without facial recognition. Our technology has applications in numerous unique use cases in sectors as diverse as retail, restaurants, airports & universities.
In addition to helping businesses better their operational efficiency while maintaining their customers' privacy, our technology can help them administer COVID-19 regulations by monitoring social distancing and other precautions via existing video feeds and presenting data and alerts.
Our proposed platform will use machine learning to understand video data similarly to other forms of conventional data by turning visual data into searchable and trackable entities.
The process can be broken down into four main steps:
- Video Feed Decoding & Storage
Our system starts with raw stream data from various sources (IP Cameras, NVRs, and Servers), which it then compresses, and structurally feeds to the cloud to allow access by the camera and room type. This step standardizes the inputs for the next steps.
- Detect, Track, and Reidentify Data
Custom trained, state of the art machine-learning-based, object detectors, and trackers allow for items (and people) in the video feed to be recognized and studied across time within a frame and across frames through reidentification techniques.
- Data Attribution
Robust classifiers extract various attributes from video images. Possible examples include age, gender, color, the presence of bags, and type of clothing. It can also classify interactions and actions being performed.
- Data Labeling
Lastly, the system stores the detected individuals, objects, and attributes as facts containing timestamped records for insight extraction, accessibility, and plotting.
For more information, contact VisionX.