Revisiting Human Action Recognition: Personalization vs. Generalization

@inproceedings{Zunino2016RevisitingHA,
  title={Revisiting Human Action Recognition: Personalization vs. Generalization},
  author={Andrea Zunino and Jacopo Cavazza and Vittorio Murino},
  booktitle={ICIAP},
  year={2016}
}
By thoroughly revisiting the classic human action recognition paradigm, we analyzed different training/testing strategies, discovering that standard (cross-validating) testing strategies are not always the suitable validation procedures to assess an algorithm’s performance. As a consequence, we design a novel action recognition architecture, applying a “personalized” strategy to learn how any subject performs any action. We discover that it is advantageous to customize (i.e., personalize) the… 

Predicting Intentions from Motion: The Subject-Adversarial Adaptation Approach

This paper addresses the problem of recognizing future actions, indeed human intentions, underlying a same initial (and apparently unrelated) motor act, and investigates how much the intention discriminants generalize across subjects, discovering that each subject tends to affect the prediction by his/her own bias.

Reference Method for the Development of Domain Action Recognition Classifiers: The Case of Medical Consultations

A reference method that assists non-experts in building classifiers for domain action recognition is proposed and it is shown that the method can be applied to a specific activity context and that the resulting classifier has an acceptable prediction accuracy.

Predicting Human Intentions from Motion Only: A 2D+3D Fusion Approach

A new multi-modal dataset consisting of a set of motion capture marker 3D data and 2D video sequences is introduced, where, by only analysing very similar movements in both training and test phases, the underlying intent can be forecast by looking at the kinematics of the immediately preceding movement.

Intention from Motion

This paper proposes Intention from Motion, a new paradigm for action prediction where, without using any contextual information, human intentions all originating from the same motor act, non specific of the following performed action, and designs a proof of concept consisting in a new multi-modal dataset.

Personalized Models in Human Activity Recognition using Deep Learning

The aim of this work is to put together deep learning techniques with incremental learning in order to obtain personalized models that perform better with respect to user-independent model and personalized model obtained using traditional machine learning techniques.

Personalized Deep Learning in Human Activity Recognition from Inertial Signals: a Preliminary Study on its Effectiveness (short paper)

Whether personalization applied to deep learning techniques can lead to more accurate models with respect to those obtained both by applying personalization to machine learning models, and to traditional deep learning models is investigated.

Personalization in Human Activity Recognition

The possibility of exploiting physical characteristics and signal similarity to achieve better results with respect to deep learning classifiers that do not rely on this information is explored.

On the Personalization of Classification Models for Human Activity Recognition

The experiments show that the employment of personalization models improves, on average, the accuracy of machine learning algorithms, thus confirming the soundness of the approach and paving the way for future investigations on this topic.

Personalizing Human-Agent Interaction Through Cognitive Models

This article outlines how it expects cognitive modeling to improve comprehension of individual cognitive processes in human-agent interaction and, particularly, human-robot interaction and proposes a conceptual framework for designing cognitive models that aims to identify whether the use of cognitive modeling is applicable to a given research question.

Beyond Covariance: SICE and Kernel Based Visual Feature Representation

To better deal with the issues of small number of feature vectors and high feature dimensionality, it is proposed to exploit the structure sparsity of visual features and exemplify sparse inverse covariance estimate as a new feature representation and to effectively model complicated feature relationship, to directly compute kernel matrix over feature dimensions.

References

SHOWING 1-10 OF 27 REFERENCES

Sequence of the Most Informative Joints (SMIJ): A new representation for human skeletal action recognition

Gesture Recognition Portfolios for Personalization

This paper addresses the problem of personalization in the context of gesture recognition, and proposes a novel and extremely efficient way of doing personalization that learns a set of classifiers during training, one of which is selected for each test subject based on the personalization data.

Personalizing Gesture Recognition Using Hierarchical Bayesian Neural Networks

This work develops hierarchical Bayesian neural networks to capture subject-specific variations and share statistical strength across subjects and builds fast, scalable algorithms for inferring the posterior distribution over all network weights in the hierarchy.

Human Action Recognition Using a Temporal Hierarchy of Covariance Descriptors on 3D Joint Locations

A novel approach to human action recognition from 3D skeleton sequences extracted from depth data that uses the covariance matrix for skeleton joint locations over time as a discriminative descriptor for a sequence to encode the relationship between joint movement and time.

The Moving Pose: An Efficient 3D Kinematics Descriptor for Low-Latency Action Recognition and Detection

A fast, simple, yet powerful non-parametric Moving Pose (MP) framework that enables low-latency recognition, one-shot learning, and action detection in difficult unsegmented sequences and is real-time, scalable, and outperforms more sophisticated approaches on challenging benchmarks.

View invariant human action recognition using histograms of 3D joints

This paper presents a novel approach for human action recognition with histograms of 3D joint locations (HOJ3D) as a compact representation of postures and achieves superior results on the challenging 3D action dataset.

A survey on vision-based human action recognition

  • R. Poppe
  • Computer Science
    Image Vis. Comput.
  • 2010

Kernelized covariance for action recognition

This paper presents Kernelized-COV, which generalizes the original covariance representation without compromising the efficiency of the computation, and validates the proposed framework against many previous approaches in the literature.