Multi-scale label interaction to improve object detection and segmentation

The project aims at advancing the state of the art in the use of context information for classification of image and video data. The use of context in the classification of images has been showed of great importance to improve the performance of actual object recognition systems. In our project we proposed the concept of Multi-scale Feature Labels as a general and compact method to exploit the local and global context. The feature extraction from the discriminative probability or classification confidence label field is of great novelty. Moreover the use of a multi-scale representation of the feature labels lead to a compact and efficient description of the context. The goal of the project has been also to provide a general-purpose method and prove its suitability in different image/video analysis problem. The two-year project generated 5 journal publications (plus 2 under submission), 10 conference publications (plus 2 under submission) and one patent (plus 1 pending). Of these publications, a relevant number make use of the main result of this project to improve the results in detection and/or segmentation of objects.

Atlas-based segmentation of pathological brain MR images

We propose a method for brain atlas deformation inpresence of large space-occupying tumors, based on an apriori model of lesion growth that assumes radialexpansion of the lesion from its starting point. First,an affine registration brings the atlas and the patientinto global correspondence. Then, the seeding of asynthetic tumor into the brain atlas provides a templatefor the lesion. Finally, the seeded atlas is deformed,combining a method derived from optical flow principlesand a model of lesion growth (MLG). Results show that themethod can be applied to the automatic segmentation ofstructures and substructures in brains with grossdeformation, with important medical applications inneurosurgery, radiosurgery and radiotherapy.

A semantic model of selective dissemination of information for digital libraries

In this paper we present the theoretical and methodologicalfoundations for the development of a multi-agentSelective Dissemination of Information (SDI) servicemodel that applies Semantic Web technologies for specializeddigital libraries. These technologies make possibleachieving more efficient information management,improving agent–user communication processes, andfacilitating accurate access to relevant resources. Othertools used are fuzzy linguistic modelling techniques(which make possible easing the interaction betweenusers and system) and natural language processing(NLP) techniques for semiautomatic thesaurus generation.Also, RSS feeds are used as “current awareness bulletins”to generate personalized bibliographic alerts.

Automated Segmentation of Cerebral Vasculature with Aneurysms in 3DRA and TOF-MRA using Geodesic Active Regions: an Evaluation Study

Purpose: To evaluate the suitability of an improved version of an automatic segmentation method based on geodesic active regions (GAR) for segmenting cerebral vasculature with aneurysms from 3D X-ray reconstruc-tion angiography (3DRA) and time of °ight magnetic resonance angiography (TOF-MRA) images available in the clinical routine.Methods: Three aspects of the GAR method have been improved: execution time, robustness to variability in imaging protocols and robustness to variability in image spatial resolutions. The improved GAR was retrospectively evaluated on images from patients containing intracranial aneurysms in the area of the Circle of Willis and imaged with two modalities: 3DRA and TOF-MRA. Images were obtained from two clinical centers, each using di®erent imaging equipment. Evaluation included qualitative and quantitative analyses ofthe segmentation results on 20 images from 10 patients. The gold standard was built from 660 cross-sections (33 per image) of vessels and aneurysms, manually measured by interventional neuroradiologists. GAR has also been compared to an interactive segmentation method: iso-intensity surface extraction (ISE). In addition, since patients had been imaged with the two modalities, we performed an inter-modality agreement analysis with respect to both the manual measurements and each of the two segmentation methods. Results: Both GAR and ISE di®ered from the gold standard within acceptable limits compared to the imaging resolution. GAR (ISE, respectively) had an average accuracy of 0.20 (0.24) mm for 3DRA and 0.27 (0.30) mm for TOF-MRA, and had a repeatability of 0.05 (0.20) mm. Compared to ISE, GAR had a lower qualitative error in the vessel region and a lower quantitative error in the aneurysm region. The repeatabilityof GAR was superior to manual measurements and ISE. The inter-modality agreement was similar between GAR and the manual measurements. Conclusions: The improved GAR method outperformed ISE qualitatively as well as quantitatively and is suitable for segmenting 3DRA and TOF-MRA images from clinical routine.

Definitional verbal patterns for semantic relation extraction

In this paper we present a description of the role of definitional verbal patterns for the extraction of semantic relations. Several studies show that semantic relations can be extracted from analytic definitions contained in machine-readable dictionaries (MRDs). In addition, definitions found in specialised texts are a good starting point to search for different types of definitions where other semantic relations occur. The extraction of definitional knowledge from specialised corpora represents another interesting approach for the extraction of semantic relations. Here, we present a descriptive analysis of definitional verbal patterns in Spanish and the first steps towards the development of a system for the automatic extraction of definitional knowledge.

A method for extracting semantic information from on-line art music discussion forums

In this paper a method for extracting semantic informationfrom online music discussion forums is proposed. The semantic relations are inferred from the co-occurrence of musical concepts in forum posts, using network analysis. The method starts by defining a dictionary of common music terms in an art music tradition. Then, it creates a complex network representation of the online forum by matchingsuch dictionary against the forum posts. Once the complex network is built we can study different network measures, including node relevance, node co-occurrence andterm relations via semantically connecting words. Moreover, we can detect communities of concepts inside the forum posts. The rationale is that some music terms are more related to each other than to other terms. All in all, this methodology allows us to obtain meaningful and relevantinformation from forum discussions.

Using qualia information to identify lexical semantic classes in an unsupervised clustering task

Acquiring lexical information is a complex problem, typically approached by relying on a number of contexts to contribute information for classification. One of the first issues to address in this domain is the determination of such contexts. The work presented here proposes the use of automatically obtained FORMAL role descriptors as features used to draw nouns from the same lexical semantic class together in an unsupervised clustering task. We have dealt with three lexical semantic classes (HUMAN, LOCATION and EVENT) in English. The results obtained show that it is possible to discriminate between elements from different lexical semantic classes using only FORMAL role information, hence validating our initial hypothesis. Also, iterating our method accurately accounts for fine-grained distinctions within lexical classes, namely distinctions involving ambiguous expressions. Moreover, a filtering and bootstrapping strategy employed in extracting FORMAL role descriptors proved to minimize effects of sparse data and noise in our task.

Automatic Lexical Semantic Classification of Nouns

The work we present here addresses cue-based noun classification in English and Spanish. Its main objective is to automatically acquire lexical semantic information by classifying nouns into previously known noun lexical classes. This is achieved by using particular aspects of linguistic contexts as cues that identify a specific lexical class. Here we concentrate on the task of identifying such cues and the theoretical background that allows for an assessment of the complexity of the task. The results show that, despite of the a-priori complexity of the task, cue-based classification is a useful tool in the automatic acquisition of lexical semantic classes.

Semantic web adoption: online tools for web evaluation and metadata extraction

This work briefly analyses the difficulties to adopt the Semantic Web, and in particular proposes systems to know the present level of migration to the different technologies that make up the Semantic Web. It focuses on the presentation and description of two tools, DigiDocSpider and DigiDocMetaEdit, designed with the aim of verifYing, evaluating, and promoting its implementation.

Frame detection over the semantic web

In the past, research in ontology learning from text has mainly focused on entity recognition, taxonomy induction and relation extraction. In this work we approach a challenging research issue: detecting semantic frames from texts and using them to encode web ontologies. We exploit a new generation Natural Language Processing technology for frame detection, and we enrich the frames acquired so far with argument restrictions provided by a super-sense tagger and domain specializations. The results are encoded according to a Linguistic MetaModel, which allows a complete translation of lexical resources and data acquired from text, enabling custom transformations of the enriched frames into modular ontology components.

Consumers’ Demand for Pork Quality: Applying Semantic Network Analysis, May 2006

Abstract Consideration of consumers’ demand for food quality entails several aspects. Quality itself is a complex and dynamic concept, and constantly evolving technical progress may cause changes in consumers’ judgment of quality. To improve our understanding of the factors influencing the demand for quality, food quality must be defined and measured from the consumer’s perspective (Cardello, 1995). The present analysis addresses the issue of food quality, focusing on pork—the food that respondents were concerned about. To gain insight into consumers’ demand, we analyzed their perception and evaluation and focused on their cognitive structures concerning pork quality. In order to more fully account for consumers’ concerns about the origin of pork, in 2004 we conducted a consumer survey of private households. The qualitative approach of concept mapping was used to uncover the cognitive structures. Network analysis was applied to interpret the results. In order to make recommendations to enterprises, we needed to know what kind of demand emerges from the given food quality schema. By establishing the importance and relative positions of the attributes, we find that the country of origin and butcher may be the two factors that have the biggest influence on consumers’ decisions about the purchase of pork.

MP2RAGE, a self bias-field corrected sequence for improved segmentation and T1-mapping at high field.

The large spatial inhomogeneity in transmit B(1) field (B(1)(+)) observable in human MR images at high static magnetic fields (B(0)) severely impairs image quality. To overcome this effect in brain T(1)-weighted images, the MPRAGE sequence was modified to generate two different images at different inversion times, MP2RAGE. By combining the two images in a novel fashion, it was possible to create T(1)-weighted images where the result image was free of proton density contrast, T(2) contrast, reception bias field, and, to first order, transmit field inhomogeneity. MP2RAGE sequence parameters were optimized using Bloch equations to maximize contrast-to-noise ratio per unit of time between brain tissues and minimize the effect of B(1)(+) variations through space. Images of high anatomical quality and excellent brain tissue differentiation suitable for applications such as segmentation and voxel-based morphometry were obtained at 3 and 7 T. From such T(1)-weighted images, acquired within 12 min, high-resolution 3D T(1) maps were routinely calculated at 7 T with sub-millimeter voxel resolution (0.65-0.85 mm isotropic). T(1) maps were validated in phantom experiments. In humans, the T(1) values obtained at 7 T were 1.15+/-0.06 s for white matter (WM) and 1.92+/-0.16 s for grey matter (GM), in good agreement with literature values obtained at lower spatial resolution. At 3 T, where whole-brain acquisitions with 1 mm isotropic voxels were acquired in 8 min, the T(1) values obtained (0.81+/-0.03 s for WM and 1.35+/-0.05 for GM) were once again found to be in very good agreement with values in the literature.

Multiparametric brainstem segmentation using a modified multivariate mixture of Gaussians

The human brainstem is a densely packed, complex but highly organised structure. It not only serves as a conduit for long projecting axons conveying motor and sensory information, but also is the location of multiple primary nuclei that control or modulate a vast array of functions, including homeostasis, consciousness, locomotion, and reflexive and emotive behaviours. Despite its importance, both in understanding normal brain function as well as neurodegenerative processes, it remains a sparsely studied structure in the neuroimaging literature. In part, this is due to the difficulties in imaging the internal architecture of the brainstem in vivo in a reliable and repeatable fashion. A modified multivariate mixture of Gaussians (mmMoG) was applied to the problem of multichannel tissue segmentation. By using quantitative magnetisation transfer and proton density maps acquired at 3 T with 0.8 mm isotropic resolution, tissue probability maps for four distinct tissue classes within the human brainstem were created. These were compared against an ex vivo fixated human brain, imaged at 0.5 mm, with excellent anatomical correspondence. These probability maps were used within SPM8 to create accurate individual subject segmentations, which were then used for further quantitative analysis. As an example, brainstem asymmetries were assessed across 34 right-handed individuals using voxel based morphometry (VBM) and tensor based morphometry (TBM), demonstrating highly significant differences within localised regions that corresponded to motor and vocalisation networks. This method may have important implications for future research into MRI biomarkers of pre-clinical neurodegenerative diseases such as Parkinson's disease.