Face segregation and recognition by cortical multi-scale line and edge coding

Models of visual perception are based on image representations in cortical area V1 and higher areas which contain many cell layers for feature extraction. Basic simple, complex and end-stopped cells provide input for line, edge and keypoint detection. In this paper we present an improved method for multi-scale line/edge detection based on simple and complex cells. We illustrate the line/edge representation for object reconstruction, and we present models for multi-scale face (object) segregation and recognition that can be embedded into feedforward dorsal and ventral data streams (the “what” and “where” subsystems) with feedback streams from higher areas for obtaining translation, rotation and scale invariance.

Face recognition by cortical multi-scale line and edge representations

Empirical studies concerning face recognition suggest that faces may be stored in memory by a few canonical representations. Models of visual perception are based on image representations in cortical area V1 and beyond, which contain many cell layers for feature extraction. Simple, complex and end-stopped cells provide input for line, edge and keypoint detection. Detected events provide a rich, multi-scale object representation, and this representation can be stored in memory in order to identify objects. In this paper, the above context is applied to face recognition. The multi-scale line/edge representation is explored in conjunction with keypoint-based saliency maps for Focus-of-Attention. Recognition rates of up to 96% were achieved by combining frontal and 3/4 views, and recognition was quite robust against partial occlusions.

Improved line/edge detection and visual reconstruction

Lines and edges provide important information for object categorization and recognition. In addition, one brightness model is based on a symbolic interpretation of the cortical multi-scale line/edge representation. In this paper we present an improved scheme for line/edge extraction from simple and complex cells and we illustrate the multi-scale representation. This representation can be used for visual reconstruction, but also for nonphotorealistic rendering. Together with keypoints and a new model of disparity estimation, a 3D wireframe representation of e.g. faces can be obtained in the future.

Cortical object segregation and categorization by multi-scale line and edge coding

In this paper we present an improved scheme for line and edge detection in cortical area V1, based on responses of simple and complex cells, truly multi-scale with no free parameters. We illustrate the multi-scale representation for visual reconstruction, and show how object segregation can be achieved with coarse-to-finescale groupings. A two-level object categorization scenario is tested in which pre-categorization is based on coarse scales only, and final categorization on coarse plus fine scales. Processing schemes are discussed in the framework of a complete cortical architecture.

Recognition of facial expressions by cortical multi-scale line and edge coding

Empirical studies concerning face recognition suggest that faces may be stored in memory by a few canonical representations. Models of visual perception are based on image representations in cortical area V1 and beyond, which contain many cell layers for feature extraction. Simple, complex and end-stopped cells provide input for line, edge and keypoint detection. Detected events provide a rich, multi-scale object representation, and this representation can be stored in memory in order to identify objects. In this paper, the above context is applied to face recognition. The multi-scale line/edge representation is explored in conjunction with keypoint-based saliency maps for Focus-of-Attention. Recognition rates of up to 96% were achieved by combining frontal and 3/4 views, and recognition was quite robust against partial occlusions.

A disparity energy model improved by line, edge and keypoint correspondences

Disparity energy models (DEMs) estimate local depth information on the basis ofVl complex cells. Our recent DEM (Martins et al, 2011 ISSPlT261-266) employs a population code. Once the population's cells have been trained with randorn-dot stereograms, it is applied at all retinotopic positions in the visual field. Despite producing good results in textured regions, the model needs to be made more precise, especially at depth transitions.

A neural network seismic detector

This experimental study focuses on a detection system at the seismic station level that should have a similar role to the detection algorithms based on the ratio STA/LTA. We tested two types of neural network: Multi-Layer Perceptrons and Support Vector Machines, trained in supervised mode. The universe of data consisted of 2903 patterns extracted from records of the PVAQ station, of the seismography network of the Institute of Meteorology of Portugal. The spectral characteristics of the records and its variation in time were reflected in the input patterns, consisting in a set of values of power spectral density in selected frequencies, extracted from a spectro gram calculated over a segment of record of pre-determined duration. The universe of data was divided, with about 60% for the training and the remainder reserved for testing and validation. To ensure that all patterns in the universe of data were within the range of variation of the training set, we used an algorithm to separate the universe of data by hyper-convex polyhedrons, determining in this manner a set of patterns that have a mandatory part of the training set. Additionally, an active learning strategy was conducted, by iteratively incorporating poorly classified cases in the training set. The best results, in terms of sensitivity and selectivity in the whole data ranged between 98% and 100%. These results compare very favorably with the ones obtained by the existing detection system, 50%.

Genetic structure at range edge: Low diversity and high inbreeding in Southeast Asian mangrove (Avicennia marina) populations

Understanding the genetic composition and mating systems of edge populations provides important insights into the environmental and demographic factors shaping species’ distribution ranges. We analysed samples of the mangrove Avicennia marina from Vietnam, northern Philippines and Australia, with microsatellite markers. We compared genetic diversity and structure in edge (Southeast Asia, and Southern Australia) and core (North and Eastern Australia) populations, and also compared our results with previously published data from core and southern edge populations. Comparisons highlighted significantly reduced gene diversity and higher genetic structure in both margins compared to core populations, which can be attributed to very low effective population size, pollinator scarcity and high environmental pressure at distribution margins. The estimated level of inbreeding was significantly higher in northeastern populations compared to core and southern populations. This suggests that despite the high genetic load usually associated with inbreeding, inbreeding or even selfing may be advantageous in margin habitats due to the possible advantages of reproductive assurance, or local adaptation. The very high level of genetic structure and inbreeding show that populations of A. marina are functioning as independent evolutionary units more than as components of a metapopulation system connected by gene flow. The combinations of those characteristics make these peripheral populations likely to develop local adaptations and therefore to be of particular interest for conservation strategies as well as for adaptation to possible future environmental changes.

On-line operation of an intelligent seismic detector

This study describes the on-line operation of a seismic detection system to act at the level of a seismic station providing similar role to that of a STA /LTA ratio-based detection algorithms. The intelligent detector is a Support Vector Machine (SVM), trained with data consisting of 2903 patterns extracted from records of the PVAQ station, one of the seismographic network's stations of the Institute of Meteorology of Portugal (IM). Records' spectral variations in time and characteristics were reflected in the SVM input patterns, as a set of values of power spectral density at selected frequencies. To ensure that all patterns of the sample data were within the range of variation of the training set, we used an algorithm to separate the universe of data by hyper-convex polyhedrons, determining in this manner a set of patterns that have a mandatory part of the training set. Additionally, an active learning strategy was conducted, by iteratively incorporating poorly classified cases in the training set. After having been trained, the proposed system was experimented in continuous operation for unseen (out of sample) data, and the SVM detector obtained 97.7% and 98.7% of sensitivity and selectivity, respectively. The same type of ANN presented 88.4 % and 99.4% of sensitivity and selectivity when applied to data of a different seismic station of IM. © 2013 Springer-Verlag Berlin Heidelberg.

BIMP: A real-time biological model of multi-scale keypoint detection in V1

We present an improved, biologically inspired and multiscale keypoint operator. Models of single- and double-stopped hypercomplex cells in area V1 of the mammalian visual cortex are used to detect stable points of high complexity at multiple scales. Keypoints represent line and edge crossings, junctions and terminations at fine scales, and blobs at coarse scales. They are detected by applying first and second derivatives to responses of complex cells in combination with two inhibition schemes to suppress responses along lines and edges. A number of optimisations make our new algorithm much faster than previous biologically inspired models, achieving real-time performance on modern GPUs and competitive speeds on CPUs. In this paper we show that the keypoints exhibit state-of-the-art repeatability in standardised benchmarks, often yielding best-in-class performance. This makes them interesting both in biological models and as a useful detector in practice. We also show that keypoints can be used as a data selection step, significantly reducing the complexity in state-of-the-art object categorisation. (C) 2014 Elsevier B.V. All rights reserved.

Reduced-Reference Image Quality Assessment Based on Edge Preservation

Assessing the subjective quality of processed images through an objective quality metric is a key issue in multimedia processing and transmission. In some scenarios, it is also important to evaluate the quality of the received images with minimal reference to the transmitted ones. For instance, for closed-loop optimisation of image and video transmission, the quality measure can be evaluated at the receiver and provided as feedback information to the system controller. The original images - prior to compression and transmission - are not usually available at the receiver side, and it is important to rely at the receiver side on an objective quality metric that does not need reference or needs minimal reference to the original images. The observation that the human eye is very sensitive to edge and contour information of an image underpins the proposal of our reduced reference (RR) quality metric, which compares edge information between the distorted and the original image. Results highlight that the metric correlates well with subjective observations, also in comparison with commonly used full-reference metrics and with a state-of-the-art reduced reference metric. © 2012 ICST Institute for Computer Science, Social Informatics and Telecommunications Engineering.

Mode III interlaminar fracture of carbon/epoxy laminates using the edge crack torsion (ECT) test

The mode III interlaminar fracture of carbon/epoxy laminates was evaluated with the edge crack torsion (ECT) test. Three-dimensional finite element analyses were performed in order to select two specimen geometries and an experimental data reduction scheme. Test results showed considerable non-linearity before the maximum load point and a significant R-curve effect. These features prevented an accurate definition of the initiation point. Nevertheless, analyses of non-linearity zones showed two likely initiation points corresponding to GIIIc values between 850 and 1100 J/m2 for both specimen geometries. Although any of these values is realistic, the range is too broad, thus showing the limitations of the ECT test and the need for further research.

Rapid automated method for on-site determination of sulfadiazine in fish farming: a stainless steel veterinary syringe coated with a selective membrane of PVC serving as a potentiometric detector in a flow-injection-analysis system

Sulfadiazine is an antibiotic of the sulfonamide group and is used as a veterinary drug in fish farming. Monitoring it in the tanks is fundamental to control the applied doses and avoid environmental dissemination. Pursuing this goal, we included a novel potentiometric design in a flow-injection assembly. The electrode body was a stainless steel needle veterinary syringe of 0.8-mm inner diameter. A selective membrane of PVC acted as a sensory surface. Its composition, the length of the electrode, and other flow variables were optimized. The best performance was obtained for sensors of 1.5-cm length and a membrane composition of 33% PVC, 66% onitrophenyloctyl ether, 1% ion exchanger, and a small amount of a cationic additive. It exhibited Nernstian slopes of 61.0 mV decade-1 down to 1.0×10-5 mol L-1, with a limit of detection of 3.1×10-6 mol L-1 in flowing media. All necessary pH/ionic strength adjustments were performed online by merging the sample plug with a buffer carrier of 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid, pH 4.9. The sensor exhibited the advantages of a fast response time (less than 15 s), long operational lifetime (60 days), and good selectivity for chloride, nitrite, acetate, tartrate, citrate, and ascorbate. The flow setup was successfully applied to the analysis of aquaculture waters. The analytical results were validated against those obtained with liquid chromatography–tandem mass spectrometry procedures. The sampling rate was about 84 samples per hour and recoveries ranged from 95.9 to 106.9%.

Data Processing at the Flaw Detector with Combined Multisector Eddy-Current Transducer

There is described data processing at the flaw detector with combined multisectional eddy-current transducer and heterofrequency magnetic field. The application of this method for detecting flaws in rods and pipes under the conditions of significant transverse displacements is described.