Regression-Based Image Alignment for General Object Categories

TL;DR

This paper introduces a regression-based approach to extend Lucas Kanade image alignment to general object categories using non-linear feature transforms like Dense SIFT, enabling robust and efficient alignment beyond facial images.

Contribution

It demonstrates how to incorporate Dense SIFT features into Lucas Kanade alignment via regression, allowing for robust alignment of diverse object categories.

Findings

Successful alignment on ImageNet objects

Extension to unsupervised joint alignment

Maintains Lucas Kanade's convergence properties

Abstract

Gradient-descent methods have exhibited fast and reliable performance for image alignment in the facial domain, but have largely been ignored by the broader vision community. They require the image function be smooth and (numerically) differentiable -- properties that hold for pixel-based representations obeying natural image statistics, but not for more general classes of non-linear feature transforms. We show that transforms such as Dense SIFT can be incorporated into a Lucas Kanade alignment framework by predicting descent directions via regression. This enables robust matching of instances from general object categories whilst maintaining desirable properties of Lucas Kanade such as the capacity to handle high-dimensional warp parametrizations and a fast rate of convergence. We present alignment results on a number of objects from ImageNet, and an extension of the method to unsupervised joint alignment of objects from a corpus of images.