Underwater target classification using wavelet packets and neural networks

Azimi-Sadjadi, M.R.; Yao, Dechen; Huang, Qiang; Dobeck, Gerald J.

doi:10.1109/72.846748

Cited by 166 publications

(100 citation statements)

References 16 publications

(3 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The training data set contained the feature vectors of backscattered data with synthesized reverberation effects with signal-to-reverberation (SRR) that corresponds to nominal operating conditions. The procedure for generating synthesized reverberation involves convolving the transmit signal with a random sequence and scaling the resultant signal according to the specified SRR [1]. The synthesized reverberation signal is then added to the backscattered signal to generate one "noisy realization."…”

Section: Resultsmentioning

confidence: 99%

“…Among these are the multivariate Gaussian classifier, the evidential K-nearest neighbor (K-NN) classifier [4], probabilistic neural network (PNN) [5], [6] and support vector machines (SVM) [7]- [9]. The performance of these systems are then compared with that of the BPNN [1] on the wideband (80 kHz) data set provided by Coastal Systems Station (CSS), Panama City, FL.…”

Section: Introductionmentioning

confidence: 99%

“…A good review of the previous methods is provided in [1]. The method in this reference uses a wavelet packet-based classification scheme to discriminate mine-like and nonmine-like objects from the acoustic backscattered signals.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Comparison of Different Classification Algorithms for Underwater Target Discrimination

Azimi-Sadjadi

Robinson

2004

IEEE Trans. Neural Netw.

Self Cite

View full text Add to dashboard Cite

Abstract-Classification of underwater targets from the acoustic backscattered signals is considered here. Several different classification algorithms are tested and benchmarked not only for their performance but also to gain insight to the properties of the feature space. Results on a wideband 80-kHz acoustic backscattered data set collected for six different objects are presented in terms of the receiver operating characteristic (ROC) and robustness of the classifiers wrt reverberation.Index Terms-K-nearest neighbor (K-NN) classifier, neural networks, probabilistic neural networks (PNNs), support vector machines (SVMs), underwater target classification.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Comparison of Different Classification Algorithms for Underwater Target Discrimination

Azimi-Sadjadi

Robinson

2004

IEEE Trans. Neural Netw.

Self Cite

View full text Add to dashboard Cite

show abstract

“…Correctly labelling underwater object is a topic which has been widely researched, especially in Mine-Counter-Measure (MCM) missions [3], [4], [5]. However the focus of those work was mainly into correctly label a portion of the sonar image, i.e.…”

Section: Related Workmentioning

confidence: 99%

On AUV actions to correctly label world information

Maurelli

Saigol

Lane

et al. 2014

2014 Oceans - St. John's

View full text Add to dashboard Cite

Abstract-An Autonomous Underwater Vehicle (AUV) needs to demonstrate a number of capabilities, in order to carry on autonomous missions with success. One of the key areas is correctly understanding the surrounding environment. However, most of the state-of-the-art approaches in labelling world information are based on the analysis of a single frame, whilst -especially in scenarios where the vehicle interact with complex structuresthere is the need of sensor data from multiple views, in order to correctly classify world information. This paper presents an active approach to solve this problem, with a tree-based (path)-planner which makes the vehicle executing a specific set of actions (following a specific trajectory), in order to discriminate among several possibilities. Results in simulation, with varying parameters, have shown that the algorithm is always bringing the robot to locations where it is expected that measurements would be different, according to the different environment.

show abstract

“…These include different kind of neural networks [30,28,31,32], genetic algorithms [33,34], decision trees [35], statistical methods such as Gaussian models and Naive Bayes [36,35], support vector machines [36,35], genetic programming [13,37,22,38], and hybrid methods [39,40,41].…”

Section: Related Workmentioning

confidence: 99%

Using Gaussian distribution to construct fitness functions in genetic programming for multiclass object classification

Zhang

Smart

2006

Pattern Recognition Letters

View full text Add to dashboard Cite

This paper describes a new approach to the use of Gaussian distribution in genetic programming (GP) for multiclass object classification problems. Instead of using predefined multiple thresholds to form different regions in the program output space for different classes, this approach uses probabilities of different classes, derived from Gaussian distributions, to construct the fitness function for classification. Two fitness measures, overlap area and weighted distribution distance, have been developed. Rather than using the best evolved program in a population, this approach uses multiple programs and a voting strategy to perform classification. The approach is examined on three multiclass object classification problems of increasing difficulty and compared with a basic GP approach. The results suggest that the new approach is more effective and more efficient than the basic GP approach. Although developed for object classification, this approach is expected to be able to be applied to other classification problems.Keywords Probability based genetic programming, object recognition, object detection, fitness function, multiclass classification Author InformationThe first author is a staff member in computer science, Victoria University of Wellington, New Zealand. The second author is postgraduate student in computer science, Victoria University of Wellington, New Zealand. Using Gaussian Distribution to Construct Fitness Functions in Genetic Programming AbstractThis paper describes a new approach to the use of Gaussian distribution in genetic programming (GP) for multiclass object classification problems. Instead of using predefined multiple thresholds to form different regions in the program output space for different classes, this approach uses probabilities of different classes, derived from Gaussian distributions, to construct the fitness function for classification. Two fitness measures, overlap area and weighted distribution distance, have been developed. Rather than using the best evolved program in a population, this approach uses multiple programs and a voting strategy to perform classification. The approach is examined on three multiclass object classification problems of increasing difficulty and compared with a basic GP approach. The results suggest that the new approach is more effective and more efficient than the basic GP approach. Although developed for object classification, this approach is expected to be able to be applied to other classification problems.

show abstract

Underwater target classification using wavelet packets and neural networks

Cited by 166 publications

References 16 publications

Comparison of Different Classification Algorithms for Underwater Target Discrimination

Comparison of Different Classification Algorithms for Underwater Target Discrimination

On AUV actions to correctly label world information

Using Gaussian distribution to construct fitness functions in genetic programming for multiclass object classification

Contact Info

Product

Resources

About