StableFlow: a physics inspired digital video stabilization

Abstract

This thesis addresses the problem of digital video stabilization. With the widespread use of handheld devices and unmanned aerial vehicles (UAVs) that has the ability to record videos, digital video stabilization becomes more important as the videos are often shaky undermining the visual quality of the video. Digital video stabilization has been studied for decades yielding an extensive amount of literature in the field, however, current approaches suffer from either being computationally expensive or under-performing in terms of visual quality . In this thesis, we firstly introduce a novel study of the effect of image denoising on feature-based digital video stabilization. Then, we introduce SteadyFlow, a novel technique for real-time stabilization inspired by the mass spring damper model. A video frame is modelled as a mass suspended in each direction by a critically dampened spring and damper which can be fine-tuned to adapt with different shaking patterns. The proposed technique is tested on video sequences that have different types of shakiness and diverse video contents. The obtained results significantly outperforms state-of-the art stabilization techniques in terms of visual quality while performing in real time.