Motion Capture of Ski Jumpers in 3 D

Motion Capture of Ski Jumpers in 3 D Trondheim University College Faculty of informatics and e-learning Ph. D student, Atle Nes Bonn, 24 -28 th of October 2004

Trondheim, Norway (summer)

Trondheim, Norway (winter)

Main research areas • Face recognition (master thesis) • Human motion analysis (current)

Scenario: Ski jumpers • Want to capture and study the motion of ski jumpers in 3 D • Results will be used to give feedback to ski jumpers that can help them to increase their jumping length

Granåsen ski jump

Capture video images • Video sequences are captured simultanuously from three video cameras • Results in large amounts of video data (about 30 MByte/sec)

Our video cameras • • AVT Marlin F 080 b (x 3) Digital IEEE 1394 Firewire 8 -bit greyscale Resolution and frame rate: 1024 x 768 x 15 fps or 640 x 480 x 30 fps

Choose feature points • Want to have automatic detection of robust feature points • Robust feature points can be human body markers (easy detectable) or naturally robust features (more difficult)

Estimate 3 D coordinates • Matching corresponding feature points from two or more cameras allows us to calculate the exact position of that feature point in 3 D (photogrammetry). • Cameras are placed such that the viewing angles give good triangulation capabilities. • Triangulation and video resolution determines the accuracy.

Track features in time • Cameras must have synchronized their video streams to ensure good 3 D coordinate accuracy when tracking moving features. • Feature localization problems with blur when object (ski jumper) is moving too fast compared to the frame rate.

Connect features back onto a 3 D model • Apply the feature motion tracks to a dynamical model of a ski jumper. • Be sure that all the movements made by the ski jumper model are allowable (cannot twist his head five times or spin his leg through the other leg). • Combine the ski jumper with a model of the ski jumping stadium.

Visualize the combined 3 D model • A CAVE environment simulating a real human view gives a much better view than just viewing the model on a regular PC screen. • The mobility of the Immersion Square is very nice.

Analyse motion • Using statistical tools • Prior knowledge about movements • Project certain movements to 2 D

Related applications • Medical: - Diagnosis of infant spontaneous movements for early detection of possible brain damage (cerebral palsy). - Diagnosis of adult movements (walk), for determination of cause of problems.

Related applications • Sports: - Study top athletes for finding optimal movement patterns. Surveillance: - Crowd surveillance and identification of possible strange behaviour in a shopping mall or airport.

Any questions?