Multi-scale cortical keypoint representation for attention and object detection

Keypoints (junctions) provide important information for focus-of-attention (FoA) and object categorization/recognition. In this paper we analyze the multi-scale keypoint representation, obtained by applying a linear and quasi-continuous scaling to an optimized model of cortical end-stopped cells, in order to study its importance and possibilities for developing a visual, cortical architecture.We show that keypoints, especially those which are stable over larger scale intervals, can provide a hierarchically structured saliency map for FoA and object recognition. In addition, the application of non-classical receptive field inhibition to keypoint detection allows to distinguish contour keypoints from texture (surface) keypoints.

Face recognition by cortical multi-scale line and edge representations

Empirical studies concerning face recognition suggest that faces may be stored in memory by a few canonical representations. Models of visual perception are based on image representations in cortical area V1 and beyond, which contain many cell layers for feature extraction. Simple, complex and end-stopped cells provide input for line, edge and keypoint detection. Detected events provide a rich, multi-scale object representation, and this representation can be stored in memory in order to identify objects. In this paper, the above context is applied to face recognition. The multi-scale line/edge representation is explored in conjunction with keypoint-based saliency maps for Focus-of-Attention. Recognition rates of up to 96% were achieved by combining frontal and 3/4 views, and recognition was quite robust against partial occlusions.

Face normalization using multi-scale cortical keypoints

Empirical studies concerning face recognition suggest that faces may be stored in memory by a few canonical representations. Models of visual perception are based on image representations in cortical area V1 and beyond, which contain many cell layers for feature extractions. Simple, complex and end-stopped cells tuned to different spatial frequencies (scales) and/or orientations provide input for line, edge and keypoint detection. This yields a rich, multi-scale object representation that can be stored in memory in order to identify objects. The multi-scale, keypoint-based saliency maps for Focus-of-Attention can be explored to obtain face detection and normalization, after which face recognition can be achieved using the line/edge representation. In this paper, we focus only on face normalization, showing that multi-scale keypoints can be used to construct canonical representations of faces in memory.

Cortical object segregation and categorization by multi-scale line and edge coding

In this paper we present an improved scheme for line and edge detection in cortical area V1, based on responses of simple and complex cells, truly multi-scale with no free parameters. We illustrate the multi-scale representation for visual reconstruction, and show how object segregation can be achieved with coarse-to-finescale groupings. A two-level object categorization scenario is tested in which pre-categorization is based on coarse scales only, and final categorization on coarse plus fine scales. Processing schemes are discussed in the framework of a complete cortical architecture.

Multi-scale keypoint hierarchy for Focus-of-Attention and object detection

Hypercolumns in area V1 contain frequency- and orientation-selective simple and complex cells for line (bar) and edge coding, plus end-stopped cells for key- point (vertex) detection. A single-scale (single-frequency) mathematical model of single and double end-stopped cells on the basis of Gabor filter responses was developed by Heitger et al. (1992 Vision Research 32 963-981). We developed an improved model by stabilising keypoint detection over neighbouring micro- scales.

Building the what and where systems: multi-scale lines, edges and keypoints

Computer vision for realtime applications requires tremendous computational power because all images must be processed from the first to the last pixel. Ac tive vision by probing specific objects on the basis of already acquired context may lead to a significant reduction of processing. This idea is based on a few concepts from our visual cortex (Rensink, Visual Cogn. 7, 17-42, 2000): (1) our physical surround can be seen as memory, i.e. there is no need to construct detailed and complete maps, (2) the bandwidth of the what and where systems is limited, i.e. only one object can be probed at any time, and (3) bottom-up, low-level feature extraction is complemented by top-down hypothesis testing, i.e. there is a rapid convergence of activities in dendritic/axonal connections.

Invariant multi-scale object categorisation and recognition

Object recognition requires that templates with canonical views are stored in memory. Such templates must somehow be normalised. In this paper we present a novel method for obtaining 2D translation, rotation and size invariance. Cortical simple, complex and end-stopped cells provide multi-scale maps of lines, edges and keypoints. These maps are combined such that objects are characterised. Dynamic routing in neighbouring neural layers allows feature maps of input objects and stored templates to converge. We illustrate the construction of group templates and the invariance method for object categorisation and recognition in the context of a cortical architecture, which can be applied in computer vision.

BIMP: A real-time biological model of multi-scale keypoint detection in V1

We present an improved, biologically inspired and multiscale keypoint operator. Models of single- and double-stopped hypercomplex cells in area V1 of the mammalian visual cortex are used to detect stable points of high complexity at multiple scales. Keypoints represent line and edge crossings, junctions and terminations at fine scales, and blobs at coarse scales. They are detected by applying first and second derivatives to responses of complex cells in combination with two inhibition schemes to suppress responses along lines and edges. A number of optimisations make our new algorithm much faster than previous biologically inspired models, achieving real-time performance on modern GPUs and competitive speeds on CPUs. In this paper we show that the keypoints exhibit state-of-the-art repeatability in standardised benchmarks, often yielding best-in-class performance. This makes them interesting both in biological models and as a useful detector in practice. We also show that keypoints can be used as a data selection step, significantly reducing the complexity in state-of-the-art object categorisation. (C) 2014 Elsevier B.V. All rights reserved.

Validation of a fast method for quantification of intra-abdominal and subcutaneous adipose tissue for large-scale human studies

Central obesity is the hallmark of a number of non-inheritable disorders. The advent of imaging techniques such asMRI has allowed for a fast and accurate assessment of body fat content and distribution. However, image analysis continues to be one of the major obstacles to the use of MRI in large-scale studies. In this study we assess the validity of the recently proposed fat–muscle quantitation system (AMRATM Profiler) for the quantification of intra-abdominal adipose tissue (IAAT) and abdominal subcutaneous adipose tissue (ASAT) from abdominal MR images. Abdominal MR images were acquired from 23 volunteers with a broad range of BMIs and analysed using sliceOmatic, the current gold-standard, and the AMRATM Profiler based on a non-rigid image registration of a library of segmented atlases. The results show that there was a highly significant correlation between the fat volumes generated by the two analysis methods, (Pearson correlation r = 0.97, p < 0.001), with the AMRATM Profiler analysis being significantly faster (~3 min) than the conventional sliceOmatic approach (~40 min). There was also excellent agreement between the methods for the quantification of IAAT (AMRA 4.73 ± 1.99 versus sliceOmatic 4.73 ± 1.75 l, p = 0.97). For the AMRATM Profiler analysis, the intra-observer coefficient of variation was 1.6% for IAAT and 1.1% for ASAT, the inter-observer coefficient of variationwas 1.4%for IAAT and 1.2%for ASAT, the intra-observer correlationwas 0.998 for IAAT and 0.999 for ASAT, and the inter-observer correlation was 0.999 for both IAAT and ASAT. These results indicate that precise and accurate measures of body fat content and distribution can be obtained in a fast and reliable form by the AMRATM Profiler, opening up the possibility of large-scale human phenotypic studies.

Comportamento do tronco na fase inicial da sequência de movimento de sentado para de pé, em indivíduos com Acidente Vascular Encefálico, em fase aguda

As alterações do controlo postural do tronco observadas em indivíduos pós acidente vascular encefálico (AVE) podem traduzir-se em dificuldades na realização da sequência de movimento de sentado para de pé (STS). Desta forma, o presente estudo teve como objetivo descrever o comportamento do tronco, durante a primeira fase do STS, em cinco indivíduos pós-AVE, em fase aguda. Para tal, os indivíduos foram avaliados em dois momentos, antes (M0) e após (M1) um período de intervenção de 3 meses. Recorreu-se ao Software de Avaliação Postural (SAPo) para analisar o deslocamento do tronco no sentido anterior, bem como a simetria entre hemitronco contralesional vs. ipsilesional. Avaliou-se também o controlo postural e equilíbrio através das Escala de Avaliação Postural para Pacientes com Sequelas de AVE (PASS) e Escala de Equilíbrio de Berg (EEB), respetivamente. Para cada indivíduo foi desenvolvido um processo de raciocínio clínico tendo por base a avaliação observacional de componentes do movimento. Os resultados evidenciaram uma diminuição do deslocamento do tronco entre M0 e M1, em três dos cinco indivíduos, no que se refere ao ângulo do plano horizontal - EIPS - ângulo inferior da omoplata e do plano horizontal – EIPS – acrómio; outro dos indivíduos, no M0 não foi capaz de realizar o deslocamento do tronco, mas concluiu-o em M1; e outro indivíduo aumentou esse deslocamento nos dois ângulos. Em relação à diferença das distâncias entre o hemitronco contralesional vs ipsilesional, para o ângulo superior da omoplata e C7, estes resultados diminuíram de M0 para M1 em todos os indivíduos. Em quatro dos cinco indivíduos, as diferenças das distâncias para o ângulo inferior da omoplata e T7 e o ângulo inferior da omoplata e a EIPS, entre o hemitronco contralesional vs ipsilesional também diminuíram no M1 comparativamente ao M0. Foi ainda possível observar que todos os indivíduos apresentaram uma melhor extensão linear do tronco na posição inicial de sentado em M1 comparativamente a M0. Para as escalas aplicadas, PASS e EEB, a maioria dos indivíduos em estudo duplicou em M1 a pontuação obtida em M0. Concluindo, os participantes do estudo evidenciaram um menor deslocamento anterior do tronco durante o STS, assim como, uma maior simetria entre os dois hemitroncos, após uma intervenção de fisioterapia baseada nos princípios do conceito de Bobath.

Adaptação para a população Portuguesa da escala de avaliação Trunk Impairment Scale (TIS)

Para uma melhor avaliação e definição do plano de intervenção do indivíduo, é cada vez mais importante a existência instrumentos de avaliação válidos e fiáveis para a população portuguesa. Objetivo: Traduzir e adaptar para a população Portuguesa a escala Trunk Impairment Scale (TIS) em pacientes pós-AVE, e avaliar as propriedades psicométricas da mesma. Metodologia: A TIS foi traduzida para o Português e adaptada culturalmente para a população portuguesa. As propriedades psicométricas da mesma, incluindo validade, fiabilidade, concordância inter-observadores, consistência interna, sensibilidade, especificidade, poder de resposta, foram avaliadas numa população diagnosticada com AVE e num grupo de controlo de participantes saudáveis. Participaram neste estudo 80 indivíduos, divididos em dois grupos, nomeadamente indivíduos pós-AVE (40) e um grupo sem patologia (40). Os participantes foram submetidos à aplicação das escalas de Berg, Medida de Independência Funcional e Escala de Desempenho Físico Fugl Meyer e a TIS de modo a avaliar as propriedades psicométricas desta. As avaliações foram realizadas por duas fisioterapeutas experientes e o re-teste foi realizado após 48 horas. Os dados foram registados e trabalhados com o programa informático SPSS 21.0. Resultados: Relativamente aos valores obtidos, verificou-se que, quanto à consistência interna da TIS estes apresentam-se de forma moderada a elevada (alfa Cronbach = 0,909). Quanto à fiabilidade inter-observadores, os itens com menor valor são os itens 1 e 4 (0,759 e 0,527, respetivamente) e os itens com valor de Kappa mais alto são os itens 5 e 6 (0,830 e 0,893, respetivamente). Relativamente à validade de critério, verificou-se que não houve correlação entre a escala de Desempenho Físico Fugl-Meyer, a escala de Equilibrio de Berg e a Medida de Independência Funcional, ou seja, os valores obtidos r=0,166; r=0,017; r= -0,002, respetivamente. Quanto à validade de construção, constatou-se que o valor da mediana é mais elevado nos itens 1 a 5, logo sugere que haja diferenças entre o grupo de indivíduos pós-AVE e o grupo de indivíduos saudáveis (p<0,001). Entre os outros dois itens (6 e 7) não foram encontradas diferenças nas respostas nos dois grupos, sendo o valor de p > 0,001. Conclusão: Os resultados obtidos neste estudo sugerem que a versão portuguesa da TIS apresenta bons níveis de fiabilidade, consistência interna e também apresenta bons resultados no que refere à concordância inter-observadores.

Assessment of hydrogen sulfide removal from biogas using a field scale anoxic biotrickling filter.

Hydrogen sulphide is one of the most toxic and corrosive compound present in swine-derived biogas streams.In this study, afield scale biotrickling filter for the removal of hydrogen sulfide was investigated.A Biofilter packed with supporting biofilm materials was fed continuously with a proprietary nutrient solution and operatedfor over 73days. The system has been operating with a H2S inlet concentrations ranging from 1,000to 3,000 ppm.Significant removal efficiencies >95% was demonstrated. pH of the stock feeding solution decreased from 6.2 to as low as 3.5within couple days.The resulting drop in pH provided circumstantial evidence to support biological H2 Soxidation to sulphuric acid by sulfide-oxidizers. Sulfur precipitation was also observed to occur. The results suggested that H2S removal from biogas stream can be efficiently achieved using portable, low cost and maintenance free biotrickling filters.

Harmonization guidelines for HLA-peptide multimer assays derived from results of a large scale international proficiency panel of the Cancer Vaccine Consortium.

PURPOSE: The Cancer Vaccine Consortium of the Cancer Research Institute (CVC-CRI) conducted a multicenter HLA-peptide multimer proficiency panel (MPP) with a group of 27 laboratories to assess the performance of the assay. EXPERIMENTAL DESIGN: Participants used commercially available HLA-peptide multimers and a well characterized common source of peripheral blood mononuclear cells (PBMC). The frequency of CD8+ T cells specific for two HLA-A2-restricted model antigens was measured by flow cytometry. The panel design allowed for participants to use their preferred staining reagents and locally established protocols for both cell labeling, data acquisition and analysis. RESULTS: We observed significant differences in both the performance characteristics of the assay and the reported frequencies of specific T cells across laboratories. These results emphasize the need to identify the critical variables important for the observed variability to allow for harmonization of the technique across institutions. CONCLUSIONS: Three key recommendations emerged that would likely reduce assay variability and thus move toward harmonizing of this assay. (1) Use of more than two colors for the staining (2) collect at least 100,000 CD8 T cells, and (3) use of a background control sample to appropriately set the analytical gates. We also provide more insight into the limitations of the assay and identified additional protocol steps that potentially impact the quality of data generated and therefore should serve as primary targets for systematic analysis in future panels. Finally, we propose initial guidelines for harmonizing assay performance which include the introduction of standard operating protocols to allow for adequate training of technical staff and auditing of test analysis procedures.