MentalEdge.ai analyzes micro-expressions using AI-driven emotional analytics to provide real-time insights. Organizations can use this information in various ways, such as enhancing training, improving team dynamics, or refining customer interactions.
The platform detects and analyzes facial micro-expressions in real time, offering insights into emotional responses during interactions, product testing, and service experiences.
MentalEdge.ai provides emotional data that organizations can use to assess workplace dynamics. Leadership teams can use this information to inform training and team development strategies.
Industries such as retail, sports, healthcare, education, and corporate training can leverage MentalEdge.ai for leadership optimization, customer engagement, and team performance insights.
MentalEdge.ai follows strict security protocols, including encryption and anonymization, to protect user data. Only authorized users can access stored information, ensuring compliance with privacy regulations.
To ensure the highest accuracy and reliability of facial recognition, it is critical that video feeds meet optimal quality standards. This document provides mandatory and recommended best practices for camera setup, environmental conditions, and capture settings.
Preferred: 4K (3840x1080)
Minimum Acceptable: 1080p (1920x1080)
Standard Applications: 30 FPS minimum (frames per second)
High-Stakes / High-Action Environments: 60 FPS strongly recommended for better motion clarity.
Use stable, lossless or lightly compressed formats where possible to prevent facial distortion. (.mp4 is preferred)
Requirement: Even, diffuse lighting without harsh shadows or glares.
Guidelines:
Use overhead lights combined with soft front lighting
Avoid strong backlighting (such as bright windows behind subjects).
In low-light environments, deploy auxiliary lighting (soft panels, LED light banks) directed towards the subjects.
Maintain at least 300–500 lux at face height for best results.
Requirement: Absolutely minimize camera shake.
Guidelines:
Always mount cameras on fixed, vibration-isolated mounts (tripods, secured mounts, wall brackets).
Avoid handheld cameras unless absolutely necessary (and only with gimbals or stabilization aids).
If using PTZ (pan-tilt-zoom) cameras, ensure movement is smooth and slow to prevent motion blur.
Optimal Positioning:
Set cameras at face height.
Angle the camera to directly face subjects when possible.
Avoid steep overhead or sharp side angles that distort facial features.
Field of View (FOV):
Frame should prioritize the face occupying approximately 20%–40% of the frame for the best analysis.
Lock exposure and lock white balance settings to avoid constant adjustment during capture.
Set ISO values low enough to minimize noise, but high enough to maintain proper exposure.
Prefer manual exposure settings where possible.
Autofocus should be fast and accurate — however, if available, locking focus after initial calibration is ideal for static setups.
Verify focus sharpness regularly.
Lower compression (higher bitrate) is preferred to preserve facial detail:
Avoid aggressive compression algorithms that introduce blurring or artifacts.
Always capture at 4K 60FPS if possible.
Minimize rapid background motion, which can interfere with tracking.
Use higher shutter speeds (e.g., 1/120s or faster) to freeze motion and reduce blur.
Use cameras with larger sensors (e.g., 1" or full-frame) if possible.
Increase lighting rather than boosting ISO excessively.
If unavoidable, prioritize capturing at 1080p with enhanced low-light settings to reduce grain and noise.
Shaky Video
Unstable mounting
Use fixed mounts; stabilize with gimbals if moving
Dark / Underexposed Feed
Poor lighting, High ISO
Add lighting sources, adjust exposure manually
Blurry Faces
Autofocus errors or motion blur
Lock focus; Increase shutter speed; ensure proper lighting
Washed-out faces
Incorrect white balance
Lock white balance to environment
Video artifacts
Overcompression
Increase bitrate; use better codecs
Delivering high-quality, consistent, and stable video is essential for optimal facial recognition performance. Following these best practices will significantly reduce errors, increase detection accuracy, and improve overall system reliability. Remember: Better input equals better recognition.
