Precision Tracking
Frames per second (FPS)
Traditional video tracking systems tend to use 8-11 FPS. Conduct Vision operates at 30 FPS, which allows for a breadth of benefits while maintaining small enough file sizes for storage efficiency.
ConductVision Model Performance
ConductVision delivers state-of-the-art accuracy across a wide range of behavioral paradigms. Our system leverages advanced AI and precision pose detection to ensure robust and reliable resultsāregardless of maze type or experimental conditions.
We evaluate model performance using industry-standard metrics, including pose mean average precision (mAP), detection precision and recall, and overall fitness score. Across mazes such as the Radial Arm, Y Maze, Morris Water Maze, and Sociability Chamber, ConductVision consistently achievesĀ over 95% accuracy, with near-perfect recall and precision in most settings.
Whether tracking exploratory behavior, decision-making, or social interaction, our models offer unparalleled reliabilityāallowing you to focus on the science, not the software.
Maze
Recall (%)
Percentage of actual positives identified
mAP50
Precision at 50% between prediction and ground truth
Fitness Score
Overall performance combining precision and recall.
ConductVision Markerless Social Interaction
Enhanced temporal resolution
Collects a more precise analysis of rapid or subtle movements that might be missed with fewer frames. Small twitches or quick transitions between behaviors can be accurately captured and analyzed.
Enhanced frame selection
Provides a larger dataset to select from for analysis. This allows for the selection of the best frames for image processing, improving analysis quality.
