Search

Early detection of the motor faults is essential and artificial neural networks are widely used for this purpose. The typical systems usually encapsulate two distinct blocks: feature extraction and classification. Such fixed and hand-crafted features may be a suboptimal choice and require a significant computational cost that will prevent their usage for real-time applications. In this paper, we propose a fast and accurate motor condition monitoring and early fault-detection system using 1-D convolutional neural networks that has an inherent adaptive design to fuse the feature extraction and classification phases of the motor fault detection into a single learning body. The proposed approach is directly applicable to the raw data (signal), and, thus, eliminates the need for a separate feature extraction algorithm resulting in more efficient systems in terms of both speed and hardware. Experimental results obtained using real motor data demonstrate the effectiveness of the proposed method for real-time motor condition monitoring. Scopus

The aim of this work was to design a personalized BCI model to detect pedaling intention through EEG signals. The approach sought to select the best among many possible BCI models for each subject. The choice was between different processing windows, feature extraction algorithms and electrode configurations. Moreover, data was analyzed offline and pseudo-online (in a way suitable for real-time applications), with a preference for the latter case. A process for selecting the best BCI model was described in detail. Results for the pseudo-online processing with the best BCI model of each subject were on average 76.7% of true positive rate, 4.94 false positives per minute and 55.1% of accuracy. The personalized BCI model approach was also found to be significantly advantageous when compared to the typical approach of using a fixed feature extraction algorithm and electrode configuration. The resulting approach could be used to more robustly interface with lower limb exoskeletons in the context of the rehabilitation of stroke patients.

A software application for automatic descriptive feature extraction from image-objects, FETEX 2.0, is presented and described in this paper. The input data include a multispectral high resolution digital image and a vector file in shapefile format containing the polygons or objects, usually extracted from a geospatial database. The design of the available descriptive features or attributes has been mainly focused on the description of agricultural parcels, providing a variety of information: spectral information from the different image bands; textural descriptors of the distribution of the intensity values based on the grey level co-occurrence matrix, the wavelet transform and a factor of edgeness; structural features describing the spatial arrangement of the elements inside the objects, based on the semivariogram curve and the Hough transform; and several descriptors of the object shape. The output file is a table that can be produced in four alternative formats, containing a vector of features for every object processed. This table of numeric values describing the objects from different points of view can be externally used as input data for any classification software. Additionally, several types of graphs and images describing the feature extraction procedure are produced, useful for interpretation and understanding the process. A test of the processing times is included, as well as an application of the program in a real parcel-based classification problem, providing some results and analyzing the applicability, the future improvement of the methodologies, and the use of additional types of data sets. This software is intended to be a dynamic tool, integrating further data and feature extraction algorithms for the progressive improvement of land use/land cover database classification and agricultural database updating processes. © 2011 Elsevier B.V., The authors appreciate the financial support provided by the Spanish Ministerio de Ciencia e Innovacion and the FEDER in the framework of the Project CGL2009-14220 and CGL2010-19591/BTE, the Spanish Institut Geografico Nacional (IGN), Institut Cartografico Valenciano (ICV), Institut Murciano de Investigacion y Desarrollo Agrario y Alimentario (IMIDA) and Banco de Terras de Galicia (Bantegal).

Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. R. P. Krish, J. Fierrez and D. Ramos, "Integrating rare minutiae in generic fingerprint matchers for forensics," 2015 IEEE International Workshop on Information Forensics and Security (WIFS), Rome, 2015, pp. 1-6. doi: 10.1109/WIFS.2015.7368557, Automated Fingerprint Identification Systems (AFIS) are commonly used by law enforcement agencies to narrow down the possible suspects from a criminal database. AFIS do not use all discriminatory features available in fingerprints but typically use only some types of features automatically extracted by a feature extraction algorithm. Latent fingerprints obtained from crime scenes are usually partial in nature which results to only very few number of reliable minutiae. Comparing a partial minutiae pattern to a full minutiae pattern is a difficult problem. Towards solving this challenge, we propose a method that exploits extended fingerprint features (unusual/rare minutiae) not commonly considered in typical minutiae-based matchers. The method we propose in this work can be combined with any existing minutiae-based matcher. We first compute a quantitative measure based on least squares between latent and tenprint minutiae points, with rare minutia feature as reference point. Then the similarity score of the reference minutiae-based matcher is modified based on the least square quantitative measure. The modified similarity score thus obtained incorporates the contribution of rare minutia features. We use a realistic forensic fingerprint casework database in our experiments which contains rare minutia features obtained from Guardia Civil, the Spanish law enforcement agency. Experiments are conducted using two reference minutiae-based matchers, namely: NIST-Bozorth3 and VeriFinger. We report a significant improvement in the rank identification accuracies when the reference minutiae matchers are augmented with our proposed algorithm based on rare minutia features., R.K. was supported by a Marie Curie Fellowship under project BBfor2 (FP7-ITN-238803) from EU. This work has also been partially supported by Spanish Guardia Civil, project Bio-Shield (TEC2012-34881) from Spanish MINECO, and project BEAT (FP7-SEC-284989) from EU.