Technology
Six technologies. One principle: AI on the device.
Everything we build runs where the camera is. Fast, offline and without data having to leave the device. And all of it is in use in our own products.

Club head detection
Our own detection model for the golf club: finds the club head frame by frame, draws the swing path and measures speed before impact. Right on the phone.

Ball detection & tracking
Finds the ball on the mat, detects launch within a few frames and reconstructs speed, launch angle and direction. The foundation of our launch monitor on a phone.

Launch monitor
Club head and ball detection, ball flight physics and 3D rendering in one app: carry, total distance, ball speed, launch angle and direction. On iPhone, iPad, Mac and projector.

Body tracking
Real-time pose estimation: joints, angles and motion phases, right on the phone. In DeepSwing it becomes a coach that splits every swing into six phases and runs thirty biomechanical checks.

Course AI & cartography
Our own segmentation model detects fairways, greens, tees, bunkers and water in aerial imagery. Combined with laser-scanned terrain and open map data it creates course maps worldwide: 40,057 golf courses captured, 10,229 of them already GPS-ready.

Edge AI & IoT
We build vision and signal AI that runs right on the device: from data and training through optimisation to the app on iOS, Android, watchOS and edge hardware.
Our stack
Our stack
- 01YOLO detection
- 02Pose estimation
- 03Semantic segmentation
- 04Core ML & Neural Engine
- 05LiteRT & NNAPI
- 06LiDAR elevation models
- 07OpenStreetMap & orthophotos
- 08Ball flight physics
- 09WebRTC
- 10Godot 3D
Contact
What should your camera recognise?
Object detection, body tracking, segmentation or cartography, on your device, in real time. Let's talk about it.