The Journal of Pattern Recognition Research (JPRR) provides an international forum for the electronic publication of high-quality research and industrial experience articles in all areas of pattern recognition, machine learning, and artificial intelligence. JPRR is committed to rigorous yet rapid reviewing. Final versions are published electronically
(ISSN 1558-884X) immediately upon acceptance.
In Press: Machine Learning Based Acoustic/IR Monitoring
JPRR Article #594
Chiropteran monitoring has become an important public concern given that wind turbines pose the threat of injury or death to bats through direct impact or barotraumas. Such monitoring therefore requires robust methodology to assess the local density, temporal activity, and diversity of bats. This work develops machine learning based monitoring approach for nocturnal ight activity of bats. It consists of Ultrasound Acoustic Monitoring System (UAMS) and thermal-Infrared Imaging Monitoring System (IIMS), the former for identi cation of bat echolocation calls and the latter for assessing flight characteristics. Supervised and unsupervised machine learning techniques were used for UAMS and IIMS, respectively. The proposed methodology was tested with real data collected during 2011 spring and fall migration around Lake Erie in Ohio. The research will be helpful for biologists and decision makers to rapidly but eff ectively assess bat density, activity, and diversity within natural areas or proposed wind development sites.
Published Online: Virtual DMA Municipal Water Supply Pipeline Leak Detection and Classification Using Advance Pattern Recognizer Multi-Class SVM
JPRR Article #548
In this paper we investigated and analyzed the concept of virtual district metered area (DMA) as the core objective of the research to resolve the current gap and limitations of the actual district metered area state of practice through the development of virtual district metered area pipeline leak detection and classification system using multi-class support vector machine (SVM) advanced pattern recognizer at Lille University water supply pipeline networks study area the so called “Zone-6”. The SVM’s were trained on multiple cases representing the presence of leaks in various sizes and locations. The research results, and analysis showed a rather promising performance, which could be successfully implemented. Moreover, the proposed method could enable the water utility companies and other stakeholders to further reduce risks associated with pipeline leaks or breaks. This method also can be used during decision-making process for selecting which pipeline requires urgent action, and engineer the optimal short-term response or alternative for maintenance strategies. Furthermore, the proposed methodology could benefit the water utility companies by reducing the cost and operational drawbacks associated with implementing the actual district metered area (DMA). It also improve the day to day operational decision making process by detecting and classifying the different stages of pipelines leaks and breaks according to their severity, which can enable the operators to see the behavior of the network on the control room screens they are familiar with and enable them to quickly perform the best short term response strategy.
In Press: Minimum Manifold-based Within-Class Scatter Support Vector Machine
JPRR Article #565
Although Support Vector Machine (SVM) is widely used in practice, it only takes the boundary information between classes into consideration while neglects the data distribution, which seriously limits the classification efficiency. In view of this, Minimum Class Variance Support Vector Machine (MCVSVM) is proposed by Zafeiriou. Compared with SVM, MCVSVM has better generalization ability because it takes both boundary information and distribution characteristics into consideration. While the above mentioned methods SVM and MCVSVM always neglect the local characteristics of each class. Based on the above analysis, this paper presents Minimum Manifold-based Within-Class Scatter Support Vector Machine (M2SVM), which not only focuses on boundary information and distribution characteristics, but also preserves the manifold structure of each class. By theory analysis, M2SVM is equivalent to SVM and MCVSVM in a certain condition. It is believed that compared with SVM and MCVSVM, M2SVM has the best generalization ability. Experiments on the man-made dataset, facial datasets and UCI datasets verify the effectiveness of the proposed method M2SVM.
Vol 9, No 1 (2014)
Pattern Recognition Theory
Speed-Up Template Matching Through Integral Image Based Weak Classifiers 1-12
Template matching is a widely used pattern recognition method, especially in industrial inspection. However, the computational costs of traditional template matching increase dramatically with both template-and scene imagesize.
JPRR Vol 9, No 1 (2014); doi:10.13176/11.516
Direct Inverse Randomized Hough Transform for Incomplete Ellipse Detection in Noisy Images 13-24
A direct inverse randomized Hough transform (DIRHT) is developed as a pre-processing procedure for incomplete ellipse detection in images with strong noise.
JPRR Vol 9, No 1 (2014); doi:10.13176/11.512
Virtual DMA Municipal Water Supply Pipeline Leak Detection and Classification Using Advance Pattern Recognizer Multi-Class SVM 25-42
In this paper we investigated and analyzed the concept of virtual DMA as the core objective of the research to resolve the current Gap and limitations of the DMA state of practice through the development of Virtual DMA Leakage Monitoring and Classification System Using Multi-class Support Vector Machine (SVM) Advanced Pattern Recognizer at Lille University WDS study area the so called “Zone-6”.
JPRR Vol 9, No 1 (2014); doi:10.13176/11.548
Vol 8, No 1 (2013)
Pattern Recognition Theory
On the Analogy of Classifier Ensembles With Primary Classifiers: Statistical Performance and Optimality 98-122
The question of how we can exploit the ability to combine different learning entities is fundamental to the core of automated pattern analysis and dictates contemporary research efforts in the field of decision fusion.
JPRR Vol 8, No 1 (2013); doi:10.13176/11.497
A System for Handwritten Script Identification From Indian Document 1-12
In a country like India a number of scripts (a total of 13) are used to write different official languages (a total of 23). For development of Optical Character Recognizer (OCR) for a particular language, the script by which the document is written is to be identified first.
JPRR Vol 8, No 1 (2013); doi:10.13176/11.485
Kernel Non-Negative Matrix Factorization for Seismic Signature Separation 13-25
A supervised learning algorithm for the separation of seismic sources in a single channel is presented. The proposed algorithm employs non-negative matrix factorization (NMF) technique in the feature space, called Kernel NMF (KNMF).
JPRR Vol 8, No 1 (2013); doi:10.13176/11.463
Stereo Disparity and Optical Flow Fusion by Geometric Relationship and an Efficient Recursive Algorithm 26-38
We suggest a relationship, called stereo-motion equation, between stereo disparity and optical flow, and a recursive filter, as an efficient algorithm to estimate the two quantities.
JPRR Vol 8, No 1 (2013); doi:10.13176/11.210
Capped K-NN Editing in Definition Lacking Environments 39-58
While any input may be contributing, imprecise specification of class of data subdivided into classes identifies as rather common a source of noise.
JPRR Vol 8, No 1 (2013); doi:10.13176/11.465
CUES: A New Hierarchical Approach for Document Clustering 66-84
Objective of the document clustering techniques is to assemble similar documents and segregate dissimilar documents. Unlike document classification, no labeled documents are provided in document clustering.
JPRR Vol 8, No 1 (2013); doi:10.13176/11.459
An Image Visual Quality Assessment Method Based on SIFT Features 85-97
There are a number of distortions in image acquisition, processing, compression, storage, transmission, and reproduction. Existing image metrics provide a good judgement on these image distortions.
JPRR Vol 8, No 1 (2013); doi:10.13176/11.511
Dismount Detection Using Kernel Sparse Representation 123-131
This paper describes a new application of kernel sparse detection technique for hyperspectral imagery (HSI) in conjunction with a sequential forward feature selection (SFFS) scheme that reduces the dimensionality to a tractable volume.
JPRR Vol 8, No 1 (2013); doi:10.13176/11.528
Short Letters
Some Features of the Users' Activities in the Mobile Telephone Network 59-65
JPRR Vol 8, No 1 (2013); doi:10.13176/11.517