Francisco Vargas

Signature verification using local directional pattern (LDP)

A method for Off-line handwritten signature verification is described in this paper. Recently, several papers have proposed pseudo dynamic methods based on the ink deposition process to discriminate between genuine and fake signatures. The major problem of those methods is the ink texture normalization in order to make the system invariable to the pen. The more extreme pen normalization is the binarization. This paper explores the usefulness of texture based measures with binarized signatures. In particularly, it proposes to apply the gray scale features local binary pattern (LBP) and local directional pattern (LDP) features to characterize black and white static signatures. The experiments done with MCYT75, GPDS300signature and GPDS960signature corpus shown that LDP are very adequate parameters for automatic verification of black and white static signatures. The results are obtained training a Support Vector Machine (SVM) classifier with genuine samples and random forgeries while random and skilled forgeries have been used for testing it.

Towards automatic wild animal monitoring: Identification of animal species in camera-trap images using very deep convolutional neural networks

Non intrusive monitoring of animals in the wild is possible using camera trapping framework, which uses cameras triggered by sensors to take a burst of images of animals in their habitat. However camera trapping framework produces a high volume of data (in the order on thousands or millions of images), which must be analyzed by a human expert. In this work, a method for animal species identification in the wild using very deep convolutional neural networks is presented. Multiple versions of the Snapshot Serengeti dataset were used in order to probe the ability of the method to cope with different challenges that camera-trap images demand. The method reached 88.9% of accuracy in Top-1 and 98.1% in Top-5 in the evaluation set using a residual network topology. Also, the results show that the proposed method outperforms previous approximations and proves that recognition in camera-trap images can be automated.

BiSpectral contactless hand based biometric system

This paper presents a contactless hand based biometric identification system using geometric and palm features. Hand images are acquired using two commercial webcams with 1200×1600 pixel resolution which are refered to as the “IR” and “visible” webcams. The IR webcam has been modified by exchanging the IR filter with a visible filter lens and reducing the gain and exposure time to improve the hand contour extraction. The hand was illuminated using 24 infra-red LEDs and 4 white light LEDs. Images acquired from the IR webcam were binarized and the normalized widths from the index to little finger were used as features. A Least Square Support Vector Machine was then used for verification. The palm features were obtained by the Orthogonal Line Ordinal Features approach applied to the image acquired by the visible webcam. The hand image from the visible webcam was segmented using an Active Shape Model guided by the hand contour from the IR webcam as an initial guest. A Hamming distance was used as verifier. More than 8000 hand images from three public databases were used in order to compare the features extraction approaches. A score level fusion of both biometrics is performed obtaining an Equal Error Rate of 0.17% with a proprietary database of 100 users acquired with the proposed device.

Angular Contour Parameterization for Signature Identification

This present work presents a parameterization system based on angles from signature edge (2D-shape) for off-line signature identification. We have used three different classifiers, the Nearest Neighbor classifier (K-NN), Neural Networks (NN) and Hidden Markov Models (HMM). Our off-line database has 800 writers with 24 samples per each writer; in total, 19200 images have been used in our experiments. We have got a success rate of 84.64%, applying as classifier Hidden Markov Model, and only used the information from this edge detection method.

Simplified EEG inverse solution for BCI real-time implementation

EEG brain imaging has become a promising approach in Brain-computer interface applications. However, accurate reconstruction of active regions and computational burden are still open issues. In this paper, we propose to use a simplified forward model that includes the reduction of the cortical dipoles based on Brodmann areas together with state-of-the-art EEG brain imaging techniques. With this approach the well known Beamformers and Greedy Search inverse solutions become feasible for real-time implementation, while guaranteeing lower localization error than previous approaches used in BCI. This methodology was tested with synthetic and real EEG data from a visual attention study. Results show zero localization error in terms of active cortical regions estimation in single 1 s trial datasets, with a computation time of 1.1 s in a non-specialized personal computer. These results open the possibility to obtain in real-time information of active cortical regions in Brain-computer interfaces.

HMM Based Evaluation of Physical Therapy Movements Using Kinect Tracking

Recognition of human activities in videos has experienced considerable changes with the introduction of cost-effective technology that allows for the tracking of individual body parts. This has led to the development of numerous tele-health applications that aim to help patients in their recovery process. Most of these systems are based on techniques to measure the degree of similarity of time series, together with thresholds to evaluate whether the movement satisfies the specification. This means that sequences similar enough to a template, but containing deviations from the correct form, may be considered correct, and thus the quality of movement incorrectly assessed. In this paper we propose the use of Hidden Markov Models as novelty detectors to evaluate the quality of movement in human beings. The results show the potential of this approach in detecting the sequences that deviate from normality for a wide range of activities common in physical therapy and rehabilitation.

Isolated interrupted aortic arch in an adult male.

A 38-year-old male was referred to our hospital for evaluation of hypertension that was diagnosed 1 year ago. The patient did not have any symptoms. On physical examination, bilateral femoral and popliteal pulses were weak, and there was an ejection systolic murmur on the left second intercostal area. Chest radiograph showed rib notching with cardiomegaly (A). Transthoracic echocardiography showed concentric left ventricular hypertrophy, and the ascending aorta was dilated. The descending aorta appeared to terminate 2 to 3 cm after the left subclavian artery in suprasternal echocardiographic view. Aortography showed a complete interruption of the aortic arch just distal to the origin of the left subclavian artery and the absence of patent ductus arteriosus. Magnetic resonance angiography confirmed the diagnosis of isolated interrupted aortic arch (IAA) with extensive collaterals (B and C). An extra-anatomic bypass procedure in a single-stage repair through a median sternotomy extended into an upper midline laparotomy was performed by placing a graft-assisted anastomosis between the ascending aorta and the supraceliac abdominal aorta. The postoperative course was uneventful. IAA is a rare congenital anomaly with an incidence of three per million live births. In most cases, IAA is associated with other congenital cardiac anomalies. The classification made by Celoria and Patton in 1959 is the most widely used, based on the position of interruption that may be distal to the left subclavian artery (type A), between the left common carotid and left subclavian arteries (type B), and between the innominate and left common carotid arteries (type C). IAA is associated with a mortality rate of more than 90% at 1 year of age if untreated. In adults, a substantial collateral circulation must be present to maintain flow and enable survival. It is also thought that, in some cases, the initial affection would be a coarctation of the aorta, which evolves to the progressive closing of the lumen. Surgery is the definitive treatment of choice. Extra-anatomic bypass procedures are the most frequently preferred method of reconstruction in adult patients with IAAs. Most reported cases in adults have been reported in a single-stage repair.

Automatic Detection of Deviations in Human Movements Using HMM: Discrete vs Continuous

Automatic detection of correct performance of movements in humans is the core of coaching and rehabilitation applications. Human movement can be studied in terms of sequential data by using different sensor technologies. This representation makes it possible to use models that use sequential data to determine if executions of a certain activity are close enough to the specification or if they must be considered to be erroneous. One of the most widely used approaches for characterization of sequential data are Hidden Markov Models (HMM). They have the advantage of being able to model processes based on data from noisy sources. In this work we explore the use of both discrete and continuous HMMs to label movement sequences as either according to a specification or deviated from it. The results show that the majority of sequences are correctly labeled by the technique, with an advantage for continuous HMM.

Two-threshold energy based fall detection using a triaxial accelerometer

Elderly fall detection based on accelerometers is an active research area. Nowadays authors are addressing specific problems such as failure rates and energy consumption, but in most cases their strategies do not conciliate these objectives. In this paper we propose a double threshold based methodology with two novel detection features, a product between the sum vector magnitude and the signal magnitude area, and a normalization of the signal magnitude area over five 1 s windows. The methodology was validated using the public Mobifall dataset, and one developed for this work. It achieved 99 % of accuracy with Mobifall, and 97 % with the self-developed dataset. This methodology is based on an activity by activity analysis performed for determining which activities are prone to fail, as an alternative way of reducing detection failures.

Fall Detection Algorithm Based on Thresholds and Residual Events

Falling is a risk factor of vital importance in elderly adults, hence, the ability to detect falls automatically is necessary to minimize the risk of injury. In this work, we develop a fall detection algorithm based in inertial sensors due its scope of activity, portability, and low cost. This algorithm detects the fall across thresholds and residual events after that occurs, for this it filters the acceleration data through three filtering methodologies and by means of the amount of acceleration difference falls from Activities of Daily Living (ADLs). The algorithm is tested in a human activity and fall dataset, showing improves respect to performance compared with algorithms detailed in the literature.