About this document
Homography Networks for Object Tracking by jawaechan is a document available to read on EtoBox.
The document presents a novel approach called Homography Decomposition Networks (HDN) for robust planar object tracking in video sequences, addressing challenges posed by rapid motion and large transformations. HDN decomposes the homography transformation into a similarity group and a residual group, significantly stabilizing the condition number of the tracking system. Extensive experiments demonstrate that HDN outperforms existing state-of-the-art planar tracking methods across multiple challenging datase
- Author
- jawaechan
- Language
- EN