Sequence Summarization Using Order-constrained Kernelized Feature Subspaces
This work addresses the challenge of capturing non-linear temporal dynamics in action recognition, which is important for applications like video analysis, but it appears incremental as it builds on existing pooling and kernel methods.
The authors tackled the problem of representing temporal evolution in multivariate time-series data for human action recognition by proposing a novel pooling method called kernelized rank pooling, which achieved state-of-the-art results on several action recognition datasets.
Representations that can compactly and effectively capture temporal evolution of semantic content are important to 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 an RKHS, 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.