Evolutionary Feature Synthesis for Object Recognition

Lin, Yingqiang; Bhanu, Bir

doi:10.1109/tsmcc.2004.841912

Cited by 71 publications

(36 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In [26], Lin and Bhanu focused on finding composite features that can be used in distinguishing objects from clutter and object identification in Synthetic Aperture Radar (SAR) images using Genetic Programming (GP). There are a multitude of features that can be extracted from SAR imagery to be used in object identification.…”

Section: Gas Used To Aid In the Design Of Radar Systemsmentioning

confidence: 99%

Radar Angle of Arrival System Design Optimization Using a Genetic Algorithm

2017

View full text Add to dashboard Cite

An approach for using a Genetic Algorithm (GA) to select radar design parameters related to beamforming and angle of arrival estimation is presented in this article. This was accomplished by first developing a simulator that could evaluate the localization performance with a given set of design parameters. The simulator output was utilized as part of the GA objective function that searched the solution space for an optimal set of design parameters. Using this approach, the authors were able to more than halve the mean squared error in degrees of the localization algorithm versus a radar design using human-selected design parameters. The results of this study indicate that this kind of approach can be used to aid in the development of an actual radar design.

show abstract

Section: Gas Used To Aid In the Design Of Radar Systemsmentioning

confidence: 99%

Radar Angle of Arrival System Design Optimization Using a Genetic Algorithm

2017

View full text Add to dashboard Cite

show abstract

“…Improved classification accuracy with fewer features was obtained compared to results obtained using the original set of primitive features in the expression recognition task. The work was expanded in [20], where co-evolutionary processing was added to the approach to enable using several sub-populations in the GP algorithm. In this case, the final FVs were formed by combining the composite features synthesized by each individual subpopulation.…”

Section: Related Workmentioning

confidence: 99%

“…More specifically, generating new FVs with high dimensionality may be laborious and time-consuming; for example, in [20], a separate subpopulation needed to be generated for each new generated feature. However, despite the increasing amount of computation required, also highdimensional representations should be considered when striving to generate efficient new FVs.…”

Section: Related Workmentioning

confidence: 99%

“…This makes the overall design and implementation of the method easier, and possibly decreases the number of adjustable parameters. Finally, in some cases (e.g., in [20] or in [24] with most of the cases), it was demonstrated that the generated features cannot always improve the final classification results, which is an issue worth taking into a deeper discussion in order to discover valid reasoning for such synthesis behavior. It could be that the search for the optimal synthesis parameters in a high-dimensional solution space gets trapped into a local optimum, ultimately yielding deficient synthesis results.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

An evolutionary feature synthesis approach for content-based audio retrieval

Mäkinen

Kıranyaz

Raitoharju

et al. 2012

J AUDIO SPEECH MUSIC PROC.

View full text Add to dashboard Cite

A vast amount of audio features have been proposed in the literature to characterize the content of audio signals. In order to overcome specific problems related to the existing features (such as lack of discriminative power), as well as to reduce the need for manual feature selection, in this article, we propose an evolutionary feature synthesis technique with a built-in feature selection scheme. The proposed synthesis process searches for optimal linear/ nonlinear operators and feature weights from a pre-defined multi-dimensional search space to generate a highly discriminative set of new (artificial) features. The evolutionary search process is based on a stochastic optimization approach in which a multi-dimensional particle swarm optimization algorithm, along with fractional global best formation and heterogeneous particle behavior techniques, is applied. Unlike many existing feature generation approaches, the dimensionality of the synthesized feature vector is also searched and optimized within a set range in order to better meet the varying requirements set by many practical applications and classifiers. The new features generated by the proposed synthesis approach are compared with typical low-level audio features in several classification and retrieval tasks. The results demonstrate a clear improvement of up to 15-20% in average retrieval performance. Moreover, the proposed synthesis technique surpasses the synthesis performance of evolutionary artificial neural networks, exhibiting a considerable capability to accurately distinguish among different audio classes.

show abstract

“…Moreover , TABLE I RELATED WORK IN VISUAL LEARNING (EC-EVOLUTIONARY COMPUTATION, GP-GENETIC PROGRAMMING, LGP-LINEAR GENETIC PROGRAMMING, CC-COOPERATIVE COEVOLUTION, NN-NEURAL NETWORK) some of the methods [11]- [13] use domain-specific knowledge (DK) and are highly specialized toward a particular application. For a comparison of the effect of domain-specific versus general knowledge in a framework of evolutionary computation for object recognition, see [15]. The approach proposed in this paper is based on evolutionary computation (EC).…”

Section: Related Work and Contributionsmentioning

confidence: 99%

Visual Learning by Coevolutionary Feature Synthesis

Krawiec

Bhanu

2005

IEEE Trans. Syst., Man, Cybern. B

Self Cite

View full text Add to dashboard Cite

Abstract-In this paper, a novel genetically inspired visual learning method is proposed. Given the training raster images, this general approach induces a sophisticated feature-based recognition system. It employs the paradigm of cooperative coevolution to handle the computational difficulty of this task. To represent the feature extraction agents, the linear genetic programming is used. The paper describes the learning algorithm and provides a firm rationale for its design. Different architectures of recognition systems are considered that employ the proposed feature synthesis method. An extensive experimental evaluation on the demanding real-world task of object recognition in synthetic aperture radar (SAR) imagery shows the ability of the proposed approach to attain high recognition performance in different operating conditions.

show abstract

Evolutionary Feature Synthesis for Object Recognition

Cited by 71 publications

References 8 publications

Radar Angle of Arrival System Design Optimization Using a Genetic Algorithm

Radar Angle of Arrival System Design Optimization Using a Genetic Algorithm

An evolutionary feature synthesis approach for content-based audio retrieval

Visual Learning by Coevolutionary Feature Synthesis

Contact Info

Product

Resources

About