Non-Linear Temporal Subspace Representations for Activity Recognition

TL;DR

This paper introduces a novel kernelized rank pooling method for representing temporal evolution in multivariate time-series data, improving human activity recognition accuracy.

Contribution

It proposes a new non-linear pooling technique using kernelized rank pooling and kernelized PCA, optimized on Grassmann manifolds for better activity recognition.

Findings

Achieves state-of-the-art results on multiple action recognition datasets.

Effectively captures non-linear temporal dynamics of actions.

Demonstrates robustness across diverse feature modalities.

Abstract

Representations that can compactly and effectively capture the temporal evolution of semantic content are important to computer vision and machine learning algorithms that operate on multi-variate time-series data. We investigate such representations motivated by the task of human action recognition. Here each data instance is encoded by a multivariate feature (such as via a deep CNN) where action dynamics are characterized by their variations in time. As these features are often non-linear, we propose a novel pooling method, kernelized rank pooling, that represents a given sequence compactly as the pre-image of the parameters of a hyperplane in a reproducing kernel Hilbert space, projections of data onto which captures their temporal order. We develop this idea further and show that such a pooling scheme can be cast as an order-constrained kernelized PCA objective. We then propose to use the parameters of a kernelized low-rank feature subspace as the representation of the sequences. We cast our formulation as an optimization problem on generalized Grassmann manifolds and then solve it efficiently using Riemannian optimization techniques. We present experiments on several action recognition datasets using diverse feature modalities and demonstrate state-of-the-art results.