Recognition Evaluation Research Articles

Security Evaluation of Negative Iris Recognition

Security Evaluation of Negative Iris Recognition

Directional dense‐trajectory‐based patterns for dynamic texture recognition

Representation of dynamic textures (DTs), well‐known as a sequence of moving textures, is a challenging problem in video analysis due to the disorientation of motion features. Analysing DTs to make them ‘understandable’ plays an important role in different applications of computer vision. In this study, an efficient approach for DT description is proposed by addressing the following novel concepts. First, the beneficial properties of dense trajectories are exploited for the first time to efficiently describe DTs instead of the whole video. Second, two substantial extensions of local vector pattern operator are introduced to form a completed model which is based on complemented components to enhance its performance in encoding directional features of motion points in a trajectory. Finally, the authors present a new framework, called directional dense trajectory patterns, which takes advantage of directional beams of dense trajectories along with spatio‐temporal features of their motion points in order to construct dense‐trajectory‐based descriptors with more robustness. Evaluations of DT recognition on different benchmark datasets (i.e. UCLA, DynTex, and DynTex++) have verified the interest of the authors’ proposal.

Using a Virtual Serious Game (Deusto-e-motion1.0) to Assess the Theory of Mind in Primary School Children: Observational Descriptive Study.

BackgroundGiven the interactive media characteristics and intrinsically motivating appeal, virtual serious games are often praised for their potential for assessment and treatment.ObjectiveThis study aims to validate and develop normative data for a virtual serious game (Deusto-e-motion1.0) for the evaluation of emotional facial expression recognition and social skills, both of which are components of the theory of mind.MethodsA total of 1236 children took part in the study. The children were classified by age (8-12 years old), gender (males=639, females=597), and educational level (between the third and sixth years of Primary Education). A total of 10 schools from the Basque Country and 20 trained evaluators participated in this study.ResultsDifferences were found in Deusto-e-motion1.0 scores between groups of children depending on age and gender. Moreover, there was a moderately significant correlation between the emotional recognition scores of Deusto-e-motion1.0 and those of the Feel facial recognition test.ConclusionsDeusto-e-motion1.0 shows concurrent validity with instruments that assess emotional recognition. Results support the adequacy of Deusto-e-motion1.0 in assessing components of the theory of mind in children.

Aminobenzosuberone derivatives as PfA-M1 inhibitors: Molecular recognition and antiplasmodial evaluation.

Aminobenzosuberone-based PfA-M1 inhibitors were explored as novel antimalarial agents against two different Plasmodium falciparum strains. The 4-phenyl derivative 7c exhibited the most encouraging growth inhibitory activity with IC50 values of 6.5-11.2µM. X-ray crystal structures and early assessment of DMPK/ADME-Tox parameters allowed us to initiate structure-based drug design approach and understand the liabilities (such as potential metabolic and aqueous solubility issues) as well as identify the opportunities for improvement of this aminobenzosuberone series. It also suggested that compound 7c should be regarded as an attractive chemical tool to investigate the different biological roles of this multifunctional PfA-M1 protein.

Evaluation of word recognition and word recall with bone conduction devices: do directional microphones free up cognitive resources?

Objective: To determine if directional microphones improve cognitive capacity in typically hearing adults. The study objectives are to evaluate differences in (1) speech recognition and (2) working memory through a word recall task between bilateral directional and omnidirectional microphone settings.Design: A conductive hearing loss was artificially induced while participants wore bilateral bone conduction hearing aids on softbands. For each hearing aid setting (bilateral omnidirectional and bilateral directional), seven blocks of seven sentences from the Hearing in Noise Test (HINT) were presented at a signal-to-noise ratio of +2 dB. Participants repeated each sentence aloud and after each block, wrote down as many of the last words as they could recall.Study sample: Thirty-five typical hearing adults and a subset (n = 20) achieving ≥80% recognition.Results: The directional microphone setting showed significant improvement over the omnidirectional setting for recognition and recall for both the full set of participants and the subset of participants with ≥ 80% recognition.Conclusions: This study demonstrated that features such as directional microphones can improve both speech recognition and working memory. Even in listening situations where participants can understand the majority of speech, directional microphones may offer improvements to cognitive capacity and reduce listening effort.

Deep optical character recognition: a case of Pashto language

Over the past decades, text recognition technologies have focused immensely on noncursive isolated scripts. A text recognition system for the cursive Pashto script will serve as a great contribution, allowing the traditional, cultural, and educational Pashto literature to be converted into machine-readable form. We propose the use of deep learning architectures based on the transfer learning for the recognition of Pashto ligatures. For recognition analysis and evaluation, the ligature images in the dataset are preprocessed by data augmentation techniques, i.e., negatives, contours, and rotated to increase the variation of each sample and size of the original dataset. Rich feature representations are automatically extracted from the Pashto ligature images using deep convolution layers of the convolution neural network (CNN) architectures using fine-tuned approach. Pretrained CNN architectures: AlexNet, GoogleNet, and VGG (VGG-16 and VGG-19) are used for classification by feeding the extracted features to a fully connected layer and a softmax layer. The proposed deep transfer-based learning has achieved phenomenal recognition rates for Pashto ligatures on benchmark FAST-NU Pashto dataset. An accuracy of 97.24%, 97.46%, and 99.03% is achieved using AlexNext, GoogleNet, and VGGNet architectures, respectively.

Arbitrary-View Human Action Recognition: A Varying-View RGB-D Action Dataset

Current researches of action recognition which focus on single-view and multi-view recognition can hardly satisfy the requirements of human-robot interaction (HRI) applications for recognizing human actions from arbitrary views. Arbitrary-view recognition is still a challenging issue due to view changes and visual occlusions. In addition, the lack of datasets also sets up barriers. To provide data for arbitrary-view action recognition, we collect a new large-scale RGB-D action dataset for arbitrary-view action analysis, including RGB videos, depth and skeleton sequences. The dataset includes action samples captured in 8 fixed viewpoints and varying-view sequences which cover the entire 360° view angles. In total, 118 persons are invited to act 40 action categories. Our dataset involves more participants, more viewpoints and a large number of samples. More importantly, it is the first dataset containing the entire 360° varying-view sequences. The dataset provides sufficient data for multi-view, cross-view and arbitrary-view action analysis. Besides, we propose a View-guided Skeleton CNN (VS-CNN) to tackle the problem of arbitrary-view action recognition. Experiment results show that the VS-CNN achieves superior performance, and our dataset provides valuable but challenging data for the evaluation of arbitrary-view recognition.

Evaluation of acoustic pattern recognition of nightingale (Luscinia megarhynchos) recordings by citizens

Acoustic pattern recognition methods introduce new perspectives for species identification, biodiversity monitoring and data validation in citizen science but are rarely evaluated in real world scenarios. In this case study we analysed the performance of a machine learning algorithm for automated bird identification to reliably identify common nightingales (Luscinia megarhynchos) in field recordings taken by users of the smartphone app Naturblick. We found that the performance of the automated identification tool was overall robust in our selected recordings. Although most of the recordings had a relatively low confidence score, a large proportion of the recordings were identified correctly.

Cosine metric learning based speaker verification

The performance of speaker verification depends on the overlap region of the decision scores of true and imposter trials. Motivated by the fact that the overlap region can be reduced by maximizing the between-class distance while minimizing the within-class variance of the trials, we present in this paper two cosine metric learning (CML) back-end algorithms. The first one, named m-CML, aims to enlarge the between-class distance with a regularization term to control the within-class variance. The second one, named v-CML, attempts to reduce the within-class variance with a regularization term to prevent the between-class distance from getting smaller. The regularization terms in the CML methods can be initialized by a traditional channel compensation method, e.g., the linear discriminant analysis. These two algorithms are combined with front-end processing for speaker verification. To validate their effectiveness, m-CML is combined with an i-vector front-end since it is good at enlarging the between-class distance of Gaussian score distributions while v-CML is combined with a d-vector or x-vector front-end as it is able to reduce the within-class variance of heavy-tailed score distributions significantly. Experimental results on the NIST and SITW speaker recognition evaluation corpora show that the proposed algorithms outperform their initialization channel compensation methods, and are competitive to the probabilistic linear discriminant analysis back-end in terms of performance. For comparison, we also applied the m-CML and v-CML methods to the i-vector and x-vector front-ends.

Enhancement of surgical hand gesture recognition using a capsule network for a contactless interface in the operating room

Background and objective:Hand gesture recognition systems in operating rooms (ORs) are crucial for browsing and controlling computer-aided devices, which have been developed to decrease the risk of contamination during surgical procedures. Methods:We proposed the use of hand gesture recognition to enhance accuracies and recognition areas with the capsule network (CapsNet) of deep neural network and Leap Motionâ Our method includes the i) extraction and preprocessing of infrared (IR) images (60 frames per second) from Leap Motion™, ii) training of various types of networks, and iii) gesture recognition evaluation in the OR. We trained the images of training dataset (N=903) and tested images (N=100) using five types of surgical hand gestures including hovering, grab, click, one peak, and two peaks by 10 subjects with various types of augmentation methods including rotate (0∘, 90∘, 180∘), scale, translation, illumination, and resize. Results: CapsNet achieved a classification accuracy of 86.46% (around 10% improvement) compared with 73.67% for the baseline convolutional neural network (CNN) and 76.4% for VGG16. Conclusions: In conclusion, the accuracy of hand gesture recognition with CapsNet was better than that of conventional CNNs, which could be used to navigate and manipulate various types of computer-aided devices and applications through contactless gesture interaction.

Das Delir beim idiopathischen Parkinson-Syndrom

The development of delirium in patients with idiopathic Parkinson's disease (IPD) is afeared complication, which is often associated with sustained worsening of motor symptoms and psychopathological sequelae. Little is known regarding the prevalence and incidence rates, course and prognosis. Clinical studies from which recommendations for evidence-based management of delirium in IPD can be derived are lacking. To summarize the state of the art regarding epidemiological and clinical features of delirium in IPD. Discussion of prevention strategies and non-pharmacological and pharmacological treatment options. A literature search was carried out in PubMed. The IPD is an independent risk factor for the development of delirium. Patients with IPD show poorer clinical outcome frequently with cognitive worsening and motor complications following development of delirium. So far no validated rating scales for recognition and course evaluation of delirium in IPD are available. Preventive strategies and non-pharmacological measures should be consistently implemented to improve management. There are insufficient data concerning pharmacotherapy with quetiapine and clozapine, whereas other neuroleptics are contraindicated for delirium in IPD due to antidopaminergic side effects.

An Upper Limb Rehabilitation Training and Evaluation System for Stroke Patients

This system combines information technology and rehabilitation medicine. It adopts Motor Imagery (MI) intervention and mental rotation training mode in order to change the traditional inefficient mode of clinical stroke rehabilitation. We developed multi-functional side recognition rehabilitation and evaluation peripheral to evaluate the rehabilitation effect of stroke patients accurately and quantitatively. The healing effect, which reveals the degree of recovery to the patients, will no longer depend on the personal experience of the rehabilitation therapist. Based on the psychological hint and a client designed with Unity 3D, it makes the treatment less boring to stimulate the patients’ initiative during the training. This system confirms that the MI Intervention can to a certain degree improve function of limb motor and sensory feedback by analyzing 38 volunteer patients’ data in Huashan Hospital and Shanghai Jing’an District Central Hospital. Precise and quantitative evaluation results are given for the further treatment.

Analytical tools for evaluation pharmaceutical treatment of neurological diseases – a case study

The paper describes implementation details of analytical method and conclusions of novel approach to clinical trials monitoring and pharmaceutical drugs evaluation. The designed tools and implemented method utilizes mobile application and neural networks based algorithms for assisting and monitoring of time-constrained symptoms reporting. The research provides proofs that strict, assisted health state reporting correlated with sensor data acquisition can objectivize clinical trials therefore making the process more accurate. The designed algorithms are applied for wearable sensor data analysis in order to evaluate the intensity of tremor or dyskinesia and further to correlate the patient’s survey data with acquired through sensors data. Based on clinical trials observations, a set of requirements for validating symptoms of neurological diseases have been formulated, concentrating on selected few, which can be registered using wearable biomedical sensors. The constructed tools utilize conventional surveying methods supplemented with biomedical sensor for neurological symptoms recognition and intensity evaluation. Developed mobile system is aimed at clinical trials assistance utilizing sensor-based state evaluation. Such quantitative approach is a supplement for patient’s subjective evaluation of health state. This work is a discussion on pros and cons of such process composition and its supplementation with technology. Existing methodology relies on patient’s health state evaluation based on iteratively answered questionnaires, which in our understanding cannot be fully controlled, thus reliable. The utilization of actigraphy and electromyography techniques provides efficient means of some gestures recognition but most of all PD tremor identification and intensity evaluation, therefore can be used for ON/OFF state and dyskinesia identification and evaluation. In order to recognize specific states for PD patients (tremors, bradykinesias, rigidity, mental slowness, etc.) a set of additional techniques have been designed and implemented. The paper contains summary of research assumptions and validation results obtained during the conceptual and testing phase of a deployed IQPharma Clinical Trials Assistant system. The original findings of presented research provide three correlating aspects (methods) of pharmaceutical therapy evaluation: 1. Patient subjective health state evaluation - based on the patient survey responses, 2. Tremor recognition and intensity evaluation – based on the inertial and biomedical sensor data analysis – aggregated features of actigraphy and electromyography 3. Reflex and mental perception assessment - based on the touch screen exercise evaluation and accuracy and providing complemental assessment of patient.

Impact of telephone cardiopulmonary resuscitation on the return of spontaneous circulation in patients with out-of-hospital cardiac arrest

The aim: Research aim is the evaluation of return of spontaneous circulation (ROSC) ratio among the individuals with out-of-hospital sudden cardiac arrest (SCA), depending on the so-called Telephone Cardiopulmonary Resuscitation (TCPR). T-CPR is based on live instruction given by the dispatcher to incident witness, who is performing CPR efforts until the EMS team arrives on scene. Material and methods: Research was based on the analysis of 782 entries (from EMS documentation) which recorded SCA and CPR. Emergency call voice recording has been rehearsed for each case, in order to confirm the capability of recognizing SCA and providing T-CPR instructions. Data was divided into two groups and the results have been compared. The ROSC ratio for both groups (“T-CPR” and “No T-CPR”) were analyzed along with the type of incident location. Results: The research has shown that 26,4% of all SCA cases researched ended up with the ROSC. In 94% of cases the dispatcher had succeeded in encouraging the witness to perform CPR with telephone instruction (T-CPR) until the EMS team has arrived. In the “T-CPR” group, 28,7% of cases have ended with ROSC. In the “No T-CPR” group, 19,7% of cases have ended with ROSC (28,7% vs. 19,7%). Conclusions: The T-CPR should be utilized by dispatcher in the form of uniform protocol. In the process of training dispatchers there should be special emphasis on the skill of recognizing SCA upon receiving a call. The evaluation of SCA recognition, T-CPR undertaken and ROSC ratio may be an effective indicator of quality monitoring within the State Emergency Medical System.

Quantitative evaluation of subjective posture recognition by physiotherapists using a 3D motion capture.

[Purpose] This study evaluated subjective posture recognition by physiotherapists with expertise in posture, examined the quantification of posture using a three-dimensional (3D) motion capture, and described posture-based characteristics. [Participants and Methods] We photographed good, normal, and bad postures in 12 participants using an infrared camera, and the resultant data were analyzed. [Results] We observed the largest displacement from a good to a bad posture in the tenth thoracic vertebra on the X-axis in the anterior–posterior direction in comparison with other index points. Further, we observed considerable differences between good and bad postures compared with other index points. Moreover, we noted significant differences between the amount of displacement between good to a normal posture and from a good to a bad posture. The vertical displacement of the Z-axis was smaller than other index points. [Conclusion] Th10 captured features from the three postures. The X-axis was displaced most between good and bad postures. Further, the amount of displacement on the Z-axis was less between good and bad posture, rendering it difficult to capture features. Therefore, the findings reported herein can be used to compare the front and rear directions of the X-axis for capturing postural changes.

Japanese Pronunciation Evaluation Based on DDNN

In recent years, speech recognition technology based on deep learning model has made great progress, and the accuracy of speech recognition has reached more than 90%. In foreign language learning, speech evaluation is an important application. Billions of foreign language learners need to practice effective pronunciation. However, due to the different goals between speech recognition and speech evaluation, a single speech recognition model cannot be directly applied to pronunciation evaluation. This paper proposes a DDNN (double-layer deep neural network) model, which includes the speech text alignment model and speech recognition model. In the first layer of the speech alignment model, a new Viterbi algorithm method is proposed to find the best path for the alignment of speech and text. In the second layer of speech evaluation and scoring, we are the first to use the CNN (Convolutional Neural Network) and RNN (Recurrent Neural Network) on the encoding part of Attention. The accuracy of CTC model reaches 76.7%, and that of attention model is 81.2%. The experimental results show that the speech and text alignment method is effective, and the speech evaluation results based on the Attention model are better. The FRR (false rejection rate), FAR (false acceptance rate), and DER (diagnostic rate) in the Attention model were 4.5%, 5.1%, and 17.9%, respectively. At the same time, the evaluation of each sentence of the DDNN model in the online experiment is within 1 second, so the model can also be applied to the online real-time evaluation of speech pronunciation.

Analysis of field uniformity and quantitative evaluation of subsurface pitting corrosion in conductors via GPEC

Structures of nonmagnetic materials are broadly used in engineering fields such as aerospace, energy, etc. Due to corrosive and hostile environments, they are vulnerable to the Subsurface Pitting Corrosion (SPC) leading to structural failure. Therefore, it is imperative to conduct periodical inspection and comprehensive evaluation of SPC using reliable nondestructive evaluation techniques. Extended from the conventional Pulsed eddy current method (PEC), Gradient-field Pulsed Eddy Current technique (GPEC) has been proposed and found to be advantageous over PEC in terms of enhanced inspection sensitivity and accuracy in evaluation and imaging of subsurface defects in nonmagnetic conductors. In this paper two GPEC probes for uniform field excitation are intensively analyzed and compared. Their capabilities in SPC evaluation and imaging are explored through simulations and experiments. The optimal position for deployment of the magnetic field sensor is determined by scrutinizing the field uniformity and inspection sensitivity to SPC based on finite element simulations. After the optimal probe structure is chosen, quantitative evaluation and imaging of SPC are investigated. Signal/image processing algorithms for SPC evaluation are proposed. Through simulations and experiments, it has been found that the T-shaped probe together with the proposed processing algorithms is advantageous and preferable for profile recognition and depth evaluation of SPC.

Profile recognition and quantitative evaluation of corrosion in coated conductors based on PMEC with the modified Theta Map

Coated conductive structures play an important role in such industrial fields as aerospace, energy, chemical engineering, etc. Because of the corrosive and hostile environments, the corrosion which is normally hidden under the conductor coating, occurs in the surface of the in-service structure. It poses a severe threat to the integrity and safety of the structure. Therefore, it is imperative to effectively detect and quantitatively evaluate the hidden corrosion via Non-destructive Evaluation (NDE) techniques. Pulse-modulation Eddy Current technique (PMEC) has been identified to be superior to other electromagnetic NDE methods, particularly regarding evaluation and imaging of corrosion. Whereas, the acquired corrosion images hardly indicate the true opening profile of the corrosion. This hinders the quantitative evaluation of the corrosion. In light of this, in this paper PMEC along with the image processing technique for imaging and evaluation of corrosion in coated conductors is intensively investigated. The image processing technique is proposed based on Theta Map (TM) which is a processing method for quantifying the source parameters of magnetic data that describe the anomalous structure. The validity of PMEC combined with modified TM for profile recognition and assessment of hidden corrosion in nonmagnetic structures is identified via experiments.

Maximal Figure-of-Merit Framework to Detect Multi-Label Phonetic Features for Spoken Language Recognition

Bottleneck features (BNFs) generated with a deep neural network (DNN) have proven to boost spoken language recognition accuracy over basic spectral features significantly. However, BNFs are commonly extracted using language-dependent tied-context phone states as learning targets. Moreover, BNFs are less phonetically expressive than the output layer in a DNN, which is usually not used as a speech feature because of its very high dimensionality hindering further post-processing. In this article, we put forth a novel deep learning framework to overcome all of the above issues and evaluate it on the 2017 NIST Language Recognition Evaluation (LRE) challenge. We use manner and place of articulation as speech attributes, which lead to low-dimensional “universal” phonetic features that can be defined across all spoken languages. To model the asynchronous nature of the speech attributes while capturing their intrinsic relationships in a given speech segment, we introduce a new training scheme for deep architectures based on a Maximal Figure of Merit (MFoM) objective. MFoM introduces non-differentiable metrics into the backpropagation-based approach, which is elegantly solved in the proposed framework. The experimental evidence collected on the recent NIST LRE 2017 challenge demonstrates the effectiveness of our solution. In fact, the performance of speech language recognition (SLR) systems based on spectral features is improved for more than 5% absolute Cavg. Finally, the F1 metric can be brought from 77.6% up to 78.1% by combining the conventional baseline phonetic BNFs with the proposed articulatory attribute features.

13 years of speaker recognition research at BUT, with longitudinal analysis of NIST SRE

In this paper, we present a brief history and a “longitudinal study” of all important milestone modelling techniques used in text independent speaker recognition since Brno University of Technology (BUT) first participated in the NIST Speaker Recognition Evaluation (SRE) in 2006—GMM MAP, GMM MAP with eigen-channel adaptation, Joint Factor Analysis, i-vector and DNN embedding (x-vector). To emphasize the historical context, the techniques are evaluated on all NIST SRE sets since 2004 on a time-machine principle, i.e. a system is always trained using all data available up till the year of evaluation. Moreover, as user-contributed audiovisual content dominates nowadays’ Internet, we representatively include the Speakers In The Wild (SITW) and VOiCES challenge datasets in the evaluation of our systems. Not only we present a comparison of the modelling techniques, but we also show the effect of sampling frequency.