ScaleNet: A Shallow Architecture for Scale Estimation
This addresses scale estimation for computer vision tasks, offering incremental improvements by integrating with existing methods.
The paper tackles the problem of estimating scale factors between images by formulating it as predicting a probability distribution and introduces ScaleNet, a shallow architecture using dilated convolutions and correlation layers, which improves performance in tasks like camera pose estimation and 3D reconstruction.
In this paper, we address the problem of estimating scale factors between images. We formulate the scale estimation problem as a prediction of a probability distribution over scale factors. We design a new architecture, ScaleNet, that exploits dilated convolutions as well as self and cross-correlation layers to predict the scale between images. We demonstrate that rectifying images with estimated scales leads to significant performance improvements for various tasks and methods. Specifically, we show how ScaleNet can be combined with sparse local features and dense correspondence networks to improve camera pose estimation, 3D reconstruction, or dense geometric matching in different benchmarks and datasets. We provide an extensive evaluation on several tasks and analyze the computational overhead of ScaleNet. The code, evaluation protocols, and trained models are publicly available at https://github.com/axelBarroso/ScaleNet.