Optimization of OmniMotion, a tracking algorithm
Tutor/a - Director/a
Favaro, Paolo
Estudiant
Farré Farrús, Martí
Tipus de document
Treball Final de Grau
Data
2025
rights
Accés obert
Editorial
Universitat Politècnica de Catalunya
UPCommons
Resum
This thesis presents Quasi-OmniFastTrack, an improved version of the OmniMotion algorithm for long-term pixel tracking in videos. The key contribution is reducing the computational expense and training time of OmniMotion while maintaining comparable tracking performance. The main bottleneck in Omni- Motion was identified to be the NeRF network used for 3D scene representation. Quasi-OmniFastTrack replaces this with a pre-trained depth estimation model, significantly reducing training time, based on the work introduced in Omni- FastTrack, hence the name. The invertible neural network for mapping between local and canonical coordinates is retained, but optimized depths are used to lift 2D pixels to 3D. Experiments show that Quasi-OmniFastTrack reduces training time by over 50% compared to OmniMotion while achieving similar qualitative tracking results on sequences with occlusions. Performance degrades somewhat on fast-moving scenes. The ablation studies demonstrate the importance of optimizing the initial depth estimates during training. While not matching OmniMotion's robustness in all scenarios, Quasi-OmniFastTrack offers a compelling speed-accuracy tradeoff, enabling long-term tracking on more videos in practical timeframes. Future work on incorporating other modifications introduced in OmniFastTrack, like long-term semantic features, could further improve tracking consistency.
Entitat col·laboradora
Universität Bern
Professorat participant
- Favaro, Paolo