Multimodal evaluation for medical image segmentation

This paper is a joint effort between five institutionsthat introduces several novel similarity measures andcombines them to carry out a multimodal segmentationevaluation. The new similarity measures proposed arebased on the location and the intensity values of themisclassified voxels as well as on the connectivity andthe boundaries of the segmented data. We showexperimentally that the combination of these measuresimprove the quality of the evaluation. The study that weshow here has been carried out using four differentsegmentation methods from four different labs applied toa MRI simulated dataset of the brain. We claim that ournew measures improve the robustness of the evaluation andprovides better understanding about the differencebetween segmentation methods.

The Multiscenario Multienvironment BioSecure Multimodal Database (BMDB)

A new multimodal biometric database designed and acquired within the framework of the European BioSecure Network of Excellence is presented. It is comprised of more than 600 individuals acquired simultaneously in three scenarios: 1) over the Internet, 2) in an office environment with desktop PC, and 3) in indoor/outdoor environments with mobile portable hardware. The three scenarios include a common part of audio/video data. Also, signature and fingerprint data have been acquired both with desktop PC and mobile portable hardware. Additionally, hand and iris data were acquired in the second scenario using desktop PC. Acquisition has been conducted by 11 European institutions. Additional features of the BioSecure Multimodal Database (BMDB) are: two acquisitionsessions, several sensors in certain modalities, balanced gender and age distributions, multimodal realistic scenarios with simple and quick tasks per modality, cross-European diversity, availability of demographic data, and compatibility with other multimodal databases. The novel acquisition conditions of the BMDB allow us to perform new challenging research and evaluation of eithermonomodal or multimodal biometric systems, as in the recent BioSecure Multimodal Evaluation campaign. A description of this campaign including baseline results of individual modalities from the new database is also given. The database is expected to beavailable for research purposes through the BioSecure Association during 2008.

The multiscenario multienvironment BioSecure Multimodal Database (BMDB)

A new multimodal biometric database designed and acquired within the framework of the European BioSecure Network of Excellence is presented. It is comprised of more than 600 individuals acquired simultaneously in three scenarios: 1) over the Internet, 2) in an office environment with desktop PC, and 3) in indoor/outdoor environments with mobile portable hardware. The three scenarios include a common part of audio/video data. Also, signature and fingerprint data have been acquired both with desktop PC and mobile portable hardware. Additionally, hand and iris data were acquired in the second scenario using desktop PC. Acquisition has been conducted by 11 European institutions. Additional features of the BioSecure Multimodal Database (BMDB) are: two acquisition sessions, several sensors in certain modalities, balanced gender and age distributions, multimodal realistic scenarios with simple and quick tasks per modality, cross-European diversity, availability of demographic data, and compatibility with other multimodal databases. The novel acquisition conditions of the BMDB allow us to perform new challenging research and evaluation of either monomodal or multimodal biometric systems, as in the recent BioSecure Multimodal Evaluation campaign. A description of this campaign including baseline results of individual modalities from the new database is also given. The database is expected to be available for research purposes through the BioSecure Association during 2008.

Multimodal target correction by local bone registration: a PET/CT evaluation

PET/CT guidance for percutaneous interventions allows biopsy of suspicious metabolically active bone lesions even when no morphological correlation is delineable in the CT images. Clinical use of PET/CT guidance with conventional step-by-step technique is time consuming and complicated especially in cases in which the target lesion is not shown in the CT image. Our recently developed multimodal instrument guidance system (IGS) for PET/CT improved this situation. Nevertheless, bone biopsies even with IGS have a trade-off between precision and intervention duration which is proportional to patient and personnel exposure to radiation. As image acquisition and reconstruction of PET may take up to 10 minutes, preferably only one time consuming combined PET/CT acquisition should be needed during an intervention. In case of required additional control images in order to check for possible patient movements/deformations, or to verify the final needle position in the target, only fast CT acquisitions should be performed. However, for precise instrument guidance accounting for patient movement and/or deformation without having a control PET image, it is essential to be able to transfer the position of the target as identified in the original PET/CT to a changed situation as shown in the control CT.

Bringing a construction site into the lab:a context-relevant lab-based evaluation of a multimodal mobile application

Mobile technology has not yet achieved widespread acceptance in the Architectural, Engineering, and Construction (AEC) industry. This paper presents work that is part of an ongoing research project focusing on the development of multimodal mobile applications for use in the AEC industry. This paper focuses specifically on a context-relevant lab-based evaluation of two input modalities – stylus and soft-keyboard v. speech-based input – for use with a mobile data collection application for concrete test technicians. The manner in which the evaluation was conducted as well as the results obtained are discussed in detail.

Evaluation Of Soft-tissue Lesions With (18)f-fdg Pet/ct: Initial Results Of A Prospective Trial.

Although MRI is utilized for planning the resection of soft-tissue tumors, it is not always capable of differentiating benign from malignant lesions. The risk of local recurrence of soft-tissue sarcomas is increased when biopsies are performed before resection and by inadequate resections. PET associated with computed tomography using fluorodeoxyglucose labeled with fluorine-18 ((18)F-FDG PET/CT) may help differentiate between benign and malignant tumors, thus avoiding inadequate resections and making prior biopsies unnecessary. The purpose of this study was to evaluate the usefulness of (18)F-FDG PET/CT in differentiating benign from malignant solid soft-tissue lesions. Patients with solid lesions of the limbs or abdominal wall detected by MRI were submitted to (18)F-FDG PET/CT. The maximum standardized uptake value (SUVmax) cutoff was determined to differentiate malignant from benign tumors. Regardless of the (18)F-FDG PET/CT results all patients underwent biopsy and surgery. MRI was performed in 54 patients, and 10 patients were excluded because of purely lipomatose or cystic lesions. (18)F-FDG PET/CT was performed in the remaining 44 patients. Histopathology revealed 26 (59%) benign and 18 (41%) malignant soft-tissue lesions. A significant difference in SUVmax was observed between benign and malignant soft-tissue lesions. The SUVmax cutoff of 3.0 differentiated malignant from benign lesions with 100% sensitivity, 83.3% specificity, 89.6% accuracy, 78.3% positive predictive value, and 100% negative predictive value. (18)F-FDG PET/CT seems to be able to differentiate benign from malignant soft-tissue lesions with good accuracy and very high negative predictive value. Incorporating (18)F-FDG PET/CT into the diagnostic algorithm of these patients may prevent inadequate resections and unnecessary biopsies.

Evaluation of distinct input methods of an intelligent wheelchair in simulated and real environments: a performance and usability study

This paper focuses on evaluating the usability of an Intelligent Wheelchair (IW) in both real and simulated environments. The wheelchair is controlled at a high-level by a flexible multimodal interface, using voice commands, facial expressions, head movements and joystick as its main inputs. A Quasi-experimental design was applied including a deterministic sample with a questionnaire that enabled to apply the System Usability Scale. The subjects were divided in two independent samples: 46 individuals performing the experiment with an Intelligent Wheelchair in a simulated environment (28 using different commands in a sequential way and 18 with the liberty to choose the command); 12 individuals performing the experiment with a real IW. The main conclusion achieved by this study is that the usability of the Intelligent Wheelchair in a real environment is higher than in the simulated environment. However there were not statistical evidences to affirm that there are differences between the real and simulated wheelchairs in terms of safety and control. Also, most of users considered the multimodal way of driving the wheelchair very practical and satisfactory. Thus, it may be concluded that the multimodal interfaces enables very easy and safe control of the IW both in simulated and real environments.

Enhancing Multimodal Silent Speech Interfaces With Feature Selection

In research on Silent Speech Interfaces (SSI), different sources of information (modalities) have been combined, aiming at obtaining better performance than the individual modalities. However, when combining these modalities, the dimensionality of the feature space rapidly increases, yielding the well-known "curse of dimensionality". As a consequence, in order to extract useful information from this data, one has to resort to feature selection (FS) techniques to lower the dimensionality of the learning space. In this paper, we assess the impact of FS techniques for silent speech data, in a dataset with 4 non-invasive and promising modalities, namely: video, depth, ultrasonic Doppler sensing, and surface electromyography. We consider two supervised (mutual information and Fisher's ratio) and two unsupervised (meanmedian and arithmetic mean geometric mean) FS filters. The evaluation was made by assessing the classification accuracy (word recognition error) of three well-known classifiers (knearest neighbors, support vector machines, and dynamic time warping). The key results of this study show that both unsupervised and supervised FS techniques improve on the classification accuracy on both individual and combined modalities. For instance, on the video component, we attain relative performance gains of 36.2% in error rates. FS is also useful as pre-processing for feature fusion. Copyright © 2014 ISCA.

A Rat Model of Diabetic Wound Infection for the Evaluation of Topical Antimicrobial Therapies

Diabetes mellitus is an epidemic multisystemic chronic disease that frequently is complicated by complex wound infections. Innovative topical antimicrobial therapy agents are potentially useful for multimodal treatment of these infections. However, an appropriately standardized in vivo model is currently not available to facilitate the screening of these emerging products and their effect on wound healing. To develop such a model, we analyzed, tested, and modified published models of wound healing. We optimized various aspects of the model, including animal species, diabetes induction method, hair removal technique, splint and dressing methods, the control of unintentional bacterial infection, sampling methods for the evaluation of bacterial burden, and aspects of the microscopic and macroscopic assessment of wound healing, all while taking into consideration animal welfare and the '3Rs' principle. We thus developed a new wound infection model in rats that is optimized for testing topical antimicrobial therapy agents. This model accurately reproduces the pathophysiology of infected diabetic wound healing and includes the current standard treatment (that is, debridement). The numerous benefits of this model include the ready availability of necessary materials, simple techniques, high reproducibility, and practicality for experiments with large sample sizes. Furthermore, given its similarities to infected-wound healing and treatment in humans, our new model can serve as a valid alternative for applied research.

Peripheral Nerve Repair: Multimodal Comparison of the Long-Term Regenerative Potential of Adipose Tissue-Derived Cells in a Biodegradable Conduit.

Tissue engineering is a popular topic in peripheral nerve repair. Combining a nerve conduit with supporting adipose-derived cells could offer an opportunity to prevent time-consuming Schwann cell culture or the use of an autograft with its donor site morbidity and eventually improve clinical outcome. The aim of this study was to provide a broad overview over promising transplantable cells under equal experimental conditions over a long-term period. A 10-mm gap in the sciatic nerve of female Sprague-Dawley rats (7 groups of 7 animals, 8 weeks old) was bridged through a biodegradable fibrin conduit filled with rat adipose-derived stem cells (rASCs), differentiated rASCs (drASCs), human (h)ASCs from the superficial and deep abdominal layer, human stromal vascular fraction (SVF), or rat Schwann cells, respectively. As a control, we resutured a nerve segment as an autograft. Long-term evaluation was carried out after 12 weeks comprising walking track, morphometric, and MRI analyses. The sciatic functional index was calculated. Cross sections of the nerve, proximal, distal, and in between the two sutures, were analyzed for re-/myelination and axon count. Gastrocnemius muscle weights were compared. MRI proved biodegradation of the conduit. Differentiated rat ASCs performed significantly better than undifferentiated rASCs with less muscle atrophy and superior functional results. Superficial hASCs supported regeneration better than deep hASCs, in line with published in vitro data. The best regeneration potential was achieved by the drASC group when compared with other adipose tissue-derived cells. Considering the ease of procedure from harvesting to transplanting, we conclude that comparison of promising cells for nerve regeneration revealed that particularly differentiated ASCs could be a clinically translatable route toward new methods to enhance peripheral nerve repair.

International comparison of the German evidence-based S3-guidelines on the diagnosis and multimodal treatment of early and locally advanced gastric cancer, including adenocarcinoma of the lower esophagus.

BACKGROUND: Clinical guidelines are essential in implementing and maintaining nationwide stage-specific diagnostic and therapeutic standards. In 2011, the first German expert consensus guideline defined the evidence for diagnosis and treatment of early and locally advanced esophagogastric cancers. Here, we compare this guideline with other national guidelines as well as current literature. METHODS: The German S3-guideline used an approved development process with de novo literature research, international guideline adaptation, or good clinical practice. Other recent evidence-based national guidelines and current references were compared with German recommendations. RESULTS: In the German S3 and other Western guidelines, adenocarcinomas of the esophagogastric junction (AEG) are classified according to formerly defined AEG I-III subgroups due to the high surgical impact. To stage local disease, computed tomography of the chest and abdomen and endosonography are reinforced. In contrast, laparoscopy is optional for staging. Mucosal cancers (T1a) should be endoscopically resected "en-bloc" to allow complete histological evaluation of lateral and basal margins. For locally advanced cancers of the stomach or esophagogastric junction (≥T3N+), preferred treatment is preoperative and postoperative chemotherapy. Preoperative radiochemotherapy is an evidence-based alternative for large AEG type I-II tumors (≥T3N+). Additionally, some experts recommend treating T2 tumors with a similar approach, mainly because pretherapeutic staging is often considered to be unreliable. CONCLUSIONS: The German S3 guideline represents an up-to-date European position with regard to diagnosis, staging, and treatment recommendations for patients with locally advanced esophagogastric cancer. Effects of perioperative chemotherapy versus chemoradiotherapy are still to be investigated for adenocarcinoma of the cardia and the lower esophagus.

Evaluation of a treatment program for individuals with an intellectual disability who have high risk behaviours

In the past two decades numerous programs have emerged to treat individuals with developmental disabilities who have sexual offending behaviours. There has, however been very few studies that systematically examine the effectiveness of long term treatment with this population. The present research examines the therapeutic outcomes of a multi-modal behaviour approach with six individuals with intellectual disabilities previously charged with sexual assault. The participants also exhibited severe behavioural challenges that included verbal aggression, physical aggression, destruction and self-injury. These six participants (5 males, 1 female) were admitted to a Long Term Residential Treatment Program (LTRTP), due to the severity of their behaviours and due to their lack of treatment success in other programs. Individualized treatment plans focused on the reduction of maladaptive behaviours and the enhancing of skills such as positive coping strategies, socio-sexual knowledge, life skills, recreation and leisure skills. The treatment program also included psychiatric, psychological, medical, behavioural and educational interventions. The participants remained in the Long Term Residential Treatment Program (LTRTP) program from 181 to 932 days (average of 1.5 years). Pre and post treatment evaluations were conducted using the following tools: frequency of target behaviours, Psychopathology Inventory for Mentally Retarded Adults (PIMRA), Emotional Problems Scale (EPS), Socio-Sexual Knowledge and Attitudes Assessment Tool (SSKAAT-R) and Quality of Life Questionnaire (QOL-Q). Recidivism rates and the need for re-hospitalization were also noted for each participant. By offering high levels of individualized interventions, all six participants showed a 37 % rate of reduction in maladaptive behaviours with zero to low rates of inappropriate sexualbehaviour, there were no psychiatric hospitalizations, and there was no recidivism for 5 of 6 participants. In addition, medication was reduced. Mental health scores on the PIMRA were reduced across all participants by 25 % and scores on the Quality of Life Questionnaire increased for all participants by an average of 72 %. These findings add to and build upon the existing literature on long term treatment benefits for individuals with a intellectual disability who sexually offend. By utilizing an individualized and multimodal treatment approach to reduce severe behavioural challenges, not only can the maladaptive behaviours be reduced, but adaptive behaviours can be increased, mental health concerns can be managed, and overall quality of life can be improved.

A rank aggregation framework for video multimodal geocoding

This paper proposes a rank aggregation framework for video multimodal geocoding. Textual and visual descriptions associated with videos are used to define ranked lists. These ranked lists are later combined, and the resulting ranked list is used to define appropriate locations for videos. An architecture that implements the proposed framework is designed. In this architecture, there are specific modules for each modality (e.g, textual and visual) that can be developed and evolved independently. Another component is a data fusion module responsible for combining seamlessly the ranked lists defined for each modality. We have validated the proposed framework in the context of the MediaEval 2012 Placing Task, whose objective is to automatically assign geographical coordinates to videos. Obtained results show how our multimodal approach improves the geocoding results when compared to methods that rely on a single modality (either textual or visual descriptors). We also show that the proposed multimodal approach yields comparable results to the best submissions to the Placing Task in 2012 using no extra information besides the available development/training data. Another contribution of this work is related to the proposal of a new effectiveness evaluation measure. The proposed measure is based on distance scores that summarize how effective a designed/tested approach is, considering its overall result for a test dataset. © 2013 Springer Science+Business Media New York.

Identificação de alunos com altas habilidades ou superdotação a partir de uma avaliação multimodal

Pós-graduação em Psicologia do Desenvolvimento e Aprendizagem - FC

Uniquitous internet middleware: architecture design and prototype evaluation

Technology advances in recent years have dramatically changed the way users exploit contents and services available on the Internet, by enforcing pervasive and mobile computing scenarios and enabling access to networked resources almost from everywhere, at anytime, and independently of the device in use. In addition, people increasingly require to customize their experience, by exploiting specific device capabilities and limitations, inherent features of the communication channel in use, and interaction paradigms that significantly differ from the traditional request/response one. So-called Ubiquitous Internet scenario calls for solutions that address many different challenges, such as device mobility, session management, content adaptation, context-awareness and the provisioning of multimodal interfaces. Moreover, new service opportunities demand simple and effective ways to integrate existing resources into new and value added applications, that can also undergo run-time modifications, according to ever-changing execution conditions. Despite service-oriented architectural models are gaining momentum to tame the increasing complexity of composing and orchestrating distributed and heterogeneous functionalities, existing solutions generally lack a unified approach and only provide support for specific Ubiquitous Internet aspects. Moreover, they usually target rather static scenarios and scarcely support the dynamic nature of pervasive access to Internet resources, that can make existing compositions soon become obsolete or inadequate, hence in need of reconfiguration. This thesis proposes a novel middleware approach to comprehensively deal with Ubiquitous Internet facets and assist in establishing innovative application scenarios. We claim that a truly viable ubiquity support infrastructure must neatly decouple distributed resources to integrate and push any kind of content-related logic outside its core layers, by keeping only management and coordination responsibilities. Furthermore, we promote an innovative, open, and dynamic resource composition model that allows to easily describe and enforce complex scenario requirements, and to suitably react to changes in the execution conditions.