C3Vision

Synergistically couple computer vision and brain-computer interface technology
to improve searching of imagery and video content

Computer Vision to EEG

C3Vision for Remote Sensing

In remote sensing and intelligence gathering, analysts search images that are several hundred giga-pixels in size in order to find and monitor objects and regions of interest. Trained analysts may need upto 60 minutes to complete review of a 30,000 × 30,000 pixel image. Can this search be enhanced to reduce analyst's time?

Neuromatters revolutionary C3Vision system leverages the high sensitivity of computer vision with the high specificity of the human visual recognition to priortitize which regions of the image should be searched first.

C3Vision Workflow

C3Vision helps analysts focus attention on regions with high target probability in five easy steps:

  1. Instantiate CV module to perform analysis of the entire image and obtain an initial ordering of regions that have high target probability
  2. Use detections to divide large image into smaller image chips for presentation via EEG RSVP. At the same time, separate set of images train the EEG classifier
  3. Classifer decodes the EEG signal generates confidence scores indicating level of interest of each image chip
  4. Scores generate a priority list ranking the chip by EEG scores
  5. Priority list fed to visualization software allowing analysts' to jump around the image based on EEG-based score

Contact us