Robust Multimodal Person Identification With Limited Training Data

This paper presents a novel method of audio-visual feature-level fusion for person identification where both the speech and facial modalities may be corrupted, and there is a lack of prior knowledge about the corruption. Furthermore, we assume there are limited amount of training data for each modality (e.g., a short training speech segment and a single training facial image for each person). A new multimodal feature representation and a modified cosine similarity are introduced to combine and compare bimodal features with limited training data, as well as vastly differing data rates and feature sizes. Optimal feature selection and multicondition training are used to reduce the mismatch between training and testing, thereby making the system robust to unknown bimodal corruption. Experiments have been carried out on a bimodal dataset created from the SPIDRE speaker recognition database and AR face recognition database with variable noise corruption of speech and occlusion in the face images. The system's speaker identification performance on the SPIDRE database, and facial identification performance on the AR database, is comparable with the literature. Combining both modalities using the new method of multimodal fusion leads to significantly improved accuracy over the unimodal systems, even when both modalities have been corrupted. The new method also shows improved identification accuracy compared with the bimodal systems based on multicondition model training or missing-feature decoding alone.

Modeling continuous self-report measures of perceived emotion using generalized additive mixed models

Emotion research has long been dominated by the “standard method” of displaying posed or acted static images of facial expressions of emotion. While this method has been useful it is unable to investigate the dynamic nature of emotion expression. Although continuous self-report traces have enabled the measurement of dynamic expressions of emotion, a consensus has not been reached on the correct statistical techniques that permit inferences to be made with such measures. We propose Generalized Additive Models and Generalized Additive Mixed Models as techniques that can account for the dynamic nature of such continuous measures. These models allow us to hold constant shared components of responses that are due to perceived emotion across time, while enabling inference concerning linear differences between groups. The mixed model GAMM approach is preferred as it can account for autocorrelation in time series data and allows emotion decoding participants to be modelled as random effects. To increase confidence in linear differences we assess the methods that address interactions between categorical variables and dynamic changes over time. In addition we provide comments on the use of Generalized Additive Models to assess the effect size of shared perceived emotion and discuss sample sizes. Finally we address additional uses, the inference of feature detection, continuous variable interactions, and measurement of ambiguity.

Validation of a passive stereophotogrammetry system for imaging of the breast: A geometric analysis

The overall aim of this study was to assess the accuracy, reproducibility and stability of a high resolution passive stereophotogrammetry system to image a female mannequin torso, to validate measurements made on the textured virtual surface compared with those obtained using manual techniques and to develop an approach to make objective measurements of the female breast. 3D surface imaging was carried out on a textured female torso and measurements made in accordance with the system of mammometrics. Linear errors in measurements were less than 0.5 mm, system calibration produced errors of less than 1.0 mm over 94% over the surface and intra-rater reliability measured by ICC = 0.999. The mean difference between manual and digital curved surface distances was 1.36 mm with maximum and minimum differences of 3.15 mm and 0.02 mm, respectively. The stereophotogrammetry system has been demonstrated to perform accurately and reliably with specific reference to breast assessment. (C) 2011 IPEM. Published by Elsevier Ltd. All rights reserved.

An anatomical and analytical study of the Modiolus: Enlightening its relevance to plastic surgery

Background: This study was designed to evaluate the structures, muscles, and fasciae of which the modiolus is composed. It can aid in the understanding and, therefore, the utilization of plastic surgery for the aesthetic or reconstructive treatment of that region, especially the angle of the mouth. Methods: Dissections of the midface were done on five different cadavers. They were of different races (3 males, 2 females). The anatomy of the modiolus was studied in detail. New anatomical observations were classified as type I through type VI. Results: The perifacial artery fascia contributed to the modiolus in four (80%) specimens and was not part of it in 1 (20%) specimen. The facial artery was anterior to it in one (20%) specimen, lateral in four (80%) specimens, and never medial to it. No significant relationship was observed between the perifacial artery fascia contribution to the modiolus and gender or race. Also, the location of the facial artery lateral or anterior to the modiolus was not significantly related to gender or race. In addition, the deep and superficial fasciae of the face converged not anterior to the masseter muscle but actually at the modiolus, which was different from observations made by others. Conclusion: The modiolus is of critical importance in aesthetic and reconstructive plastic surgery of the face. © 2008 Springer Science+Business Media, LLC and International Society of Aesthetic Plastic Surgery.

Photosensitisers - the progression from photodynamic therapy to anti-infective surfaces

Introduction: The application of light as a stimulus in pharmaceutical systems and the associated ability to provide precise spatiotemporal control over location, wavelength and intensity, allowing ease of external control independent of environmental conditionals, has led to its increased use. Of particular note is the use of light with photosensitisers.

Areas covered: Photosensitisers are widely used in photodynamic therapy to cause a cidal effect towards cells on irradiation due to the generation of reactive oxygen species. These cidal effects have also been used to treat infectious diseases. The effects and benefits of photosensitisers in the treatment of such conditions are still being developed and further realised, with the design of novel delivery strategies. This review provides an overview of the realisation of the pharmaceutically relevant uses of photosensitisers, both in the context of current research and in terms of current clinical application, and looks to the future direction of research.

Expert opinion: Substantial advances have been and are being made in the use of photosensitisers. Of particular note are their antimicrobial applications, due to absence of resistance that is so frequently associated with conventional treatments. Their potency of action and the ability to immobilise to polymeric supports is opening a wide range of possibilities with great potential for use in healthcare infection prevention strategies.

Anatomic dissection of the arterial supply of the lips:an anatomical and analytical approach

Previous investigators have not described some of the new anatomic variations or provided quantitative and analytical data of the arterial anatomy of the lips in as much depth as in this study. Dissections of 14 different facial sides of cadavers were done. Through investigating the arterial supply of the upper and lower lips, measurements were performed and statistically analyzed. The main arterial supply of the upper lip was from the superior labial artery (SLA, mean external diameter, 1.8 mm [SD, 0.74 mm]); in addition, the subalar and septal branches contributed to its vascularization. The origin of the SLA was above the labial commissure in 78.6%. The subalar branch was not found but replaced by the alar artery that arose from the infraorbital artery in 1 specimen. The main arterial supply of the lower lip was derived from 3 branches of the facial artery, the inferior labial artery (mean external diameters, 1.4 mm [SD, 0.31 mm]) and the horizontal and vertical labiomental arteries. The inferior labial artery originated mostly below the labial commissure in 42.9% and formed a common trunk with the SLA in 28.6%. The horizontal labiomental artery was present in all, but vertical labiomental artery was absent in 21.4% of specimens. Overall, observed anatomic variations were classified into types I to VIII. Significant relations between the demographic variables and measured parameters were reported including the correlation coefficient among evaluated parameters. In conclusion, this study provides various information that aids in creating new flaps and supports the vascular base for clinical procedures in reconstructive surgery of the lip.

Laughter Type Recognition from Whole Body Motion

Despite the importance of laughter in social interactions it remains little studied in affective computing. Respiratory, auditory, and facial laughter signals have been investigated but laughter-related body movements have received almost no attention. The aim of this study is twofold: first an investigation into observers' perception of laughter states (hilarious, social, awkward, fake, and non-laughter) based on body movements alone, through their categorization of avatars animated with natural and acted motion capture data. Significant differences in torso and limb movements were found between animations perceived as containing laughter and those perceived as nonlaughter. Hilarious laughter also differed from social laughter in the amount of bending of the spine, the amount of shoulder rotation and the amount of hand movement. The body movement features indicative of laughter differed between sitting and standing avatar postures. Based on the positive findings in this perceptual study, the second aim is to investigate the possibility of automatically predicting the distributions of observer's ratings for the laughter states. The findings show that the automated laughter recognition rates approach human rating levels, with the Random Forest method yielding the best performance.

Perception and Automatic Recognition of Laughter from Whole-Body Motion: Continuous and Categorical Perspectives

Despite its importance in social interactions, laughter remains little studied in affective computing. Intelligent virtual agents are often blind to users’ laughter and unable to produce convincing laughter themselves. Respiratory, auditory, and facial laughter signals have been investigated but laughter-related body movements have received less attention. The aim of this study is threefold. First, to probe human laughter perception by analyzing patterns of categorisations of natural laughter animated on a minimal avatar. Results reveal that a low dimensional space can describe perception of laughter “types”. Second, to investigate observers’ perception of laughter (hilarious, social, awkward, fake, and non-laughter) based on animated avatars generated from natural and acted motion-capture data. Significant differences in torso and limb movements are found between animations perceived as laughter and those perceived as non-laughter. Hilarious laughter also differs from social laughter. Different body movement features were indicative of laughter in sitting and standing avatar postures. Third, to investigate automatic recognition of laughter to the same level of certainty as observers’ perceptions. Results show recognition rates of the Random Forest model approach human rating levels. Classification comparisons and feature importance analyses indicate an improvement in recognition of social laughter when localized features and nonlinear models are used.

Global Sensitivity Analysis for MAP Inference in Graphical Models

We study the sensitivity of a MAP configuration of a discrete probabilistic graphical model with respect to perturbations of its parameters. These perturbations are global, in the sense that simultaneous perturbations of all the parameters (or any chosen subset of them) are allowed. Our main contribution is an exact algorithm that can check whether the MAP configuration is robust with respect to given perturbations. Its complexity is essentially the same as that of obtaining the MAP configuration itself, so it can be promptly used with minimal effort. We use our algorithm to identify the largest global perturbation that does not induce a change in the MAP configuration, and we successfully apply this robustness measure in two practical scenarios: the prediction of facial action units with posed images and the classification of multiple real public data sets. A strong correlation between the proposed robustness measure and accuracy is verified in both scenarios.

Semi-Qualitative Probabilistic Networks in Computer Vision Problems

This paper explores the application of semi-qualitative probabilistic networks (SQPNs) that combine numeric and qualitative information to computer vision problems. Our version of SQPN allows qualitative influences and imprecise probability measures using intervals. We describe an Imprecise Dirichlet model for parameter learning and an iterative algorithm for evaluating posterior probabilities, maximum a posteriori and most probable explanations. Experiments on facial expression recognition and image segmentation problems are performed using real data.

Breaking the Intimate Screen: Pre-Recording, Special Effects and the Aesthetics of Early British Television

The technological constraints of early British television encouraged drama productions which emphasised the immediate, the enclosed and the close-up, an approach which Jason Jacobs described in the title of his seminal study as 'the intimate screen'. While Jacobs' book showed that this conception of early British television drama was only part of the reality, he did not focus on the role that special effects played in expanding the scope of the early television screen. This article will focus upon this role, showing that special effects were not only of use in expanding the temporal and spatial scope of television, but were also considered to be of interest to the audience as a way of exploring the new medium, receiving coverage in the popular press. These effects included pre-recorded film inserts, pre-recorded narration, multiple sets, model work and animation, combined with the live studio performances. Drawing upon archival research into television production files and scripts as well as audience responses and periodical coverage of television at the time of broadcast, this article will focus on telefantasy. This genre offered particular opportunities for utilising effects in ways that seemed appropriate for the experimentation with the form of television and for the drama narratives. This period also saw a variety of shifts within television as the BBC sought to determine a specific identity and understand the possibilities for the new medium.
This research also incorporates the BBC's own research and internal dialogue concerning audiences and how their tastes should best be met, at a time when the television audience was not only growing in terms of number but was also expanding geographically and socially beyond the moneyed Londoners who could afford the first television sets and were within range of the Alexandra Palace transmissions. The primary case study for this article will be the 1949 production of H.G.Wells’ The Time Machine, which incorporated pre-recorded audio and film inserts, which expanded the narrative out of the live studio performance both temporally and spatially, with the effects work receiving coverage in the popular magazine Illustrated. Other productions considered will be the 1938 and 1948 productions of RUR, the 1948 production of Blithe Spirit, and the 1950 adaptation of The Strange Case of Dr Jekyll and Mr Hyde. Despite the focus on telefantasy, this article will also include examples from other genres, both dramatic and factual, showing how the BBC's response to the changing television audience was to restrict drama to a more 'realistic' aesthetic and to move experimentation with televisual form to non-drama productions such as variety performances.

Mer and fac isomerism in tris chelate diimine metal complexes

In this perspective, we highlight the issue of meridional (mer) and facial (fac) orientation of asymmetrical diimines in tris-chelate transition metal complexes. Diimine ligands have long been the workhorse of coordination chemistry, and whilst there are now good strategies to isolate materials where the inherent metal centered chirality is under almost complete control, and systematic methodologies to isolate heteroleptic complexes, the conceptually simple geometrical isomerism has not been widely investigated. In systems where the two donor atoms are significantly different in terms of the σ-donor and π-accepting ability, the fac isomer is likely to be the thermodynamic product. For the diimine complexes with two trigonal planar nitrogen atoms there is much more subtlety to the system, and external factors such as the solvent, lattice packing and the various steric considerations play a delicate role in determining the observed and isolable product. In this article we discuss the possibilities to control the isomeric ratio in labile systems, consider the opportunities to separate inert complexes and discuss the observed differences in their spectroscopic properties. Finally we report on the ligand orientation in supramolecular systems where facial coordination leads to simple regular structures such as helicates and tetrahedra, but the ability of the ligand system to adopt a mer orientation enables self-assembled structures of considerable beauty and complexity. 

Robust bimodal person identification using face and speech with limited training data and corruption of both modalities

This paper presents a novel method of audio-visual fusion for person identification where both the speech and facial modalities may be corrupted, and there is a lack of prior knowledge about the corruption. Furthermore, we assume there is a limited amount of training data for each modality (e.g., a short training speech segment and a single training facial image for each person). A new representation and a modified cosine similarity are introduced for combining and comparing bimodal features with limited training data as well as vastly differing data rates and feature sizes. Optimal feature selection and multicondition training are used to reduce the mismatch between training and testing, thereby making the system robust to unknown bimodal corruption. Experiments have been carried out on a bimodal data set created from the SPIDRE and AR databases with variable noise corruption of speech and occlusion in the face images. The new method has demonstrated improved recognition accuracy.

Nerve Growth Factor Sensitisation of TRPA1 on Sensory Neurones

Background: Sensory neurones from the trigeminal nerve innervate the oro-facial region and teeth. Transient receptor potential channels (TRPs) expressed by these neurones are responsible for relaying sensory information such as changes in ambient temperature, mechanical sensations and pain. Study of TRP channel expression and regulation in human sensory neurones therefore merits investigation to improve our understanding of allodynia and hyperalgesia. Objective: The objective of this study was to differentiate human dental pulp stem cells (hDPSCs) towards a neuronal phenotype (peripheral neuronal equivalents; PNEs) and employ this model to study TRP channel sensitisation. Method: hDPSCs were enriched by preferential adhesion to fibronectin, plated on coverslips (thickness 0) coated with poly-l-ornithine and laminin and then differentiated for 7 days in neurobasal A medium with additional supplementation. A whole cell patch clamp technique was used to investigate whether TRP channels on PNE membranes were modulated in the presence of nerve growth factor (NGF). PNEs were treated with NGF for 20 minutes immediately before experimentation and then stimulated for TRPA1 activity using cinnamaldehyde. Peak currents were read at 80 mV and -80 mV and compared to peak currents recorded in untreated PNEs. Data were analysed and plotted using Clampfit9 software (Molecular Devices, Sunnyvale, California, USA). Result: Results showed for the first time that pre-treatment of PNEs by NGF produced significantly larger inward and outward currents demonstrating that TRPA1 channels on PNE membranes were capable of becoming sensitised following treatment with NGF. Conclusion: Sensitisation of TRPA1 by NGF provides evidence of a mechanism for rapid neuronal sensitisation that is independent of TRPA1 gene expression

An in vitro model of nociceptors in human trigeminal nerves

Background: The oro-facial region is densely innervated by the trigeminal nerve, which when stimulated can induce noxious pain sensation and contribute to neurogenic inflammation in local tissues. Recent research on the expression of specialised ion channels on the trigeminal nerve has highlighted the need to undertake more extensive studies on ion channel expression/functionality with the aim of elucidating their role in pain sensations. A major family of such ion channels is the transient receptor potential (TRP) channels which are activated by a wide variety of thermal, mechanical or chemical stimuli and merit investigation as possible druggable targets for future analgesics.
Objective: Study of TRP channel expression and regulation in oro-facial tissues is hindered by the fact that the cell bodies of neurons innervating these tissues are located in the trigeminal ganglion. Using dental pulp stem cells differentiated towards peripheral neuronal equivalents (PNEs), we sought to determine TRP channel expression, functionality and potential modulation by cytokines in this novel model.
Method: Dental pulp stem cells (DPSCs) were grown on substrate-coated tissue culture plates and differentiated towards a neuronal phenotype using neuronal induction media. Quantitative polymerase chain reaction (qPCR) was performed on PNEs +/-cytokine treatment. Ion channel functionality was investigated using whole cell patch clamping.
Result: qPCR analysis showed that PNEs expressed the TRP channels TRPA1, TRPV1, TRPV4 and TRPM8. TRPA1 was the most abundantly expressed TRP channel studied whereas TRPM8 was lowly expressed. TRP channel expression was shown to be regulated by treatment with inflammatory cytokines. Patch clamp studies using specific agonists and antagonists for TRPA1 and TRPV1 showed these channels were functional.
Conclusion: PNEs differentiated from DPSCs provide a suitable model for TRP channel expression, regulation, and sensistisation in oro-facial tissues. This human neuronal model has potential for use in pre-clinical studies of novel analgesics.