Highly accurate computer vision made easy to use.

We are deep and different. We build intelligent, proprietary deep networks that outperform the industry using far less training data. We lead the world in person detection, face detection, and facial recognition. Companies that have found OpenCV to be too slow or inaccurate for their needs turn to Sighthound to add artificial intelligence for vision to their products.

Our technology is available in the cloud, on embedded chips, or on Windows, Mac and Linux computers.

Read more about our capabilities below or contact us to learn more.

Vehicle Detection

From static or moving cameras we detect, classify and track vehicles.

Face Detection

Highly accurate and more robust in real world scenarios that involve detecting small or distant faces.

Click here for benchmark accuracy results.

"Desktop Summit Group Photo" by Kat, used under CC BY 2.0 / Cropped from original

Face Recognition

The most accurate facial recognition software in the world.

Click here for benchmark accuracy results.

Gender Recognition

Software that classifies male or female with greater than 95% accuracy from a picture where the face is clearly visible.

Person Detection

From still, moving or flying cameras. Real-time person detection.

Facial Landmark Detection

Sighthound places 68 landmark dots on the face allowing the identification of features such as eyes, nose, mouth and giving pose and emotion information.