CyLKs: Unsupervised Cycle Lucas-Kanade Network for Landmark Tracking

TL;DR

CyLKs introduces an unsupervised learning framework that enhances Lucas-Kanade tracking by learning feature transformations with CNNs, reducing drift and relaxing brightness consistency assumptions for improved landmark tracking.

Contribution

It presents a novel unsupervised training method for LK tracking using CNNs to learn feature spaces, improving robustness without requiring annotations.

Findings

Improved landmark tracking accuracy on THUMOS and 300VW datasets.

Reduced drift and better handling of large motions and occlusions.

Effective unsupervised training of LK-based tracking models.

Abstract

Across a majority of modern learning-based tracking systems, expensive annotations are needed to achieve state-of-the-art performance. In contrast, the Lucas-Kanade (LK) algorithm works well without any annotation. However, LK has a strong assumption of photometric (brightness) consistency on image intensity and is easy to drift because of large motion, occlusion, and aperture problem. To relax the assumption and alleviate the drift problem, we propose CyLKs, a data-driven way of training Lucas-Kanade in an unsupervised manner. CyLKs learns a feature transformation through CNNs, transforming the input images to a feature space which is especially favorable to LK tracking. During training, we perform differentiable Lucas-Kanade forward and backward on the convolutional feature maps, and then minimize the re-projection error. During testing, we perform the LK tracking on the learned features. We apply our model to the task of landmark tracking and perform experiments on datasets of THUMOS and 300VW.