A dynamic optimization approach for adaptive incremental learning

Kapp, Marcelo N.; Sabourin, Robert; Maupin, Patrick

doi:10.1002/int.20501

Cited by 6 publications

(5 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Incremental and adaptive learning techniques have been applied rather sparingly to the field of human action recognition [28]- [31], since they are more frequent in related fields as, for instance, visual tracking [32], [33]. In incremental learning, the goal is to improve the learnt model or exemplarbased data, combining the previous experience with the knowledge extracted from the new example(s), in order to both successfully recognise new samples and also improve the recognition of existing ones.…”

Section: B Incremental and Adaptive Human Action Recognitionmentioning

confidence: 99%

“…Specifically, in incremental learning approaches, it is difficult to set the appropriate parameter values of the algorithm if the data is initially unknown. Therefore, the algorithm's configuration should dynamically adapt itself to the new requirements by tuning its configuration [28], for instance, by means of evolutionary algorithms. Ryoo et al [31] proposed a method to learn novel human activities incrementally.…”

Section: B Incremental and Adaptive Human Action Recognitionmentioning

confidence: 99%

See 1 more Smart Citation

Adaptive Human Action Recognition With an Evolving Bag of Key Poses

Chaaraoui

Flórez-Revuelta

2014

IEEE Trans. Auton. Mental Dev.

View full text Add to dashboard Cite

Abstract-Vision-based human action recognition allows to detect and understand meaningful human motion. This makes it possible to perform advanced human-computer interaction, among other applications. In dynamic environments, adaptive methods are required to support changing scenario characteristics. Specifically, in human-robot interaction, smooth interaction between humans and robots can only be performed if these are able to evolve and adapt to the changing nature of the scenarios. In this paper, an adaptive vision-based human action recognition method is proposed. By means of an evolutionary optimisation method, adaptive and incremental learning of human actions is supported. Through an evolving bag of key poses, which models the learnt actions over time, the current learning memory is developed to recognise increasingly more actions or actors. The evolutionary method selects the optimal subset of training instances, features and parameter values for each learning phase, and handles the evolution of the model. The experimentation shows that our proposal achieves to adapt to new actions or actors successfully, by rearranging the learnt model. Stable and accurate results have been obtained on four publicly available RGB and RGB-D datasets, unveiling the method's robustness and applicability.

show abstract

Section: B Incremental and Adaptive Human Action Recognitionmentioning

confidence: 99%

Section: B Incremental and Adaptive Human Action Recognitionmentioning

confidence: 99%

Adaptive Human Action Recognition With an Evolving Bag of Key Poses

Chaaraoui

Flórez-Revuelta

2014

IEEE Trans. Auton. Mental Dev.

View full text Add to dashboard Cite

show abstract

“…Step 4: If the gbest has not been changed for more than N mu iterations, the gbest will be modified by Equation 9.…”

Section: The Hybrid Algorithmsmentioning

confidence: 99%

“…In the past decades, feedforward neural networks (FNN) have been widely used to approximate arbitrary continuous functions and classify nonlinearly separable patterns . Various algorithms are used to train FNN, such as backpropagation algorithm (BP), simulating annealing algorithm (SA), genetic algorithm (GA), and particle swarm optimization algorithm (PSO) …”

Section: Introductionmentioning

confidence: 99%

“…3,4 Various algorithms are used to train FNN, such as backpropagation algorithm (BP), simulating annealing algorithm (SA), 5 genetic algorithm (GA), 6,7 and particle swarm optimization algorithm (PSO). 8,9 As a gradient descent method, BP with good local search ability is easily trapped into local minima and converges very slowly. [10][11][12] To solve these problems, many improved algorithms were proposed.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Improved Particle Swarm Optimization Combined with Backpropagation for Feedforward Neural Networks

Han

Zhu

2012

Int. J. Intell. Syst.

View full text Add to dashboard Cite

Traditional particle swarm optimization (PSO) has good global search ability, but it easily loses its diversity and thus leads to premature convergence. Gradient descent methods such as backpropagation (BP) algorithm have good performance in searching local minima, whereas they are apt to converge to local minima. To improve search ability, two hybrid algorithms combining two improved PSOs individually with BP are proposed to train single‐hidden‐layer feedforward neural networks in this paper. In the two improved PSOs, other than the phases of repulsion and attraction, a new phase named as a mixed phase is introduced, in which the particles are attracting and repelling simultaneously to prevent premature convergence. Moreover, a modified mutation operation is performed to help particles jump out of local minima in the improved PSOs. The proposed hybrid methods achieve better convergence performance with faster convergence rate than some commonly used PSO–BP approaches and purely global or local search methods. The experiments results on function approximation and benchmark classification problems are given to verify the effectiveness and efficiency of the proposed hybrid algorithms.

show abstract