AdaBoost Multiple Feature Selection and Combination for Face Recognition

Gabor features have been recognized as one of the most successful face representations. Encouraged by the results given by this approach, other kind of facial representations based on Steerable Gaussian first order kernels and Harris corner detector are proposed in this paper. In order to reduce the high dimensional feature space, PCA and LDA techniques are employed. Once the features have been extracted, AdaBoost learning algorithm is used to select and combine the most representative features. The experimental results on XM2VTS database show an encouraging recognition rate, showing an important improvement with respect to face descriptors only based on Gabor filters.

Automatic detection and classification of football players

Color segmentation of images usually requires a manual selection and classification of samples to train the system. This paper presents an automatic system that performs these tasks without the need of a long training, providing a useful tool to detect and identify figures. In real situations, it is necessary to repeat the training process if light conditions change, or if, in the same scenario, the colors of the figures and the background may have changed, being useful a fast training method. A direct application of this method is the detection and identification of football players.

A hardware solution on multi-field packet classification

In this paper, a hardware solution for packet classification based on multi-fields is presented. The proposed scheme focuses on a new architecture based on the decomposition method. A hash circuit is used in order to reduce the memory space required for the Recursive Flow Classification (RFC) algorithm. The implementation results show that the proposed architecture achieves significant performance advantage that is comparable to that of some well-known algorithms. The solution is based on Altera Stratix III FPGA technology.

Comparison of teleconsultations and face-to-face consultations:preliminary results of a United Kingdom multicentre teledermatology study

The objective of this multicentre study was to undertake a systematic comparison of face-to-face consultations and teleconsultations performed using low-cost videoconferencing equipment. One hundred and twenty-six patients were enrolled by their general practitioners across three sites. Each patient underwent a teleconsultation with a distant dermatologist followed by a traditional face-to-face consultation with a dermatologist. The main outcome measures were diagnostic concordance rates, management plans and patient and doctor satisfaction. One hundred and fifty-five diagnoses were identified by the face-to-face consultations from the sample of 126 patients. Identical diagnoses were recorded from both types of consultation in 59% of cases. Teledermatology consultations missed a secondary diagnosis in 6% of cases and were unable to make a useful diagnosis in 11% of cases. Wrong diagnoses were made by the teledermatologist in 4% of cases. Dermatologists were able to make a definitive diagnosis by face-to-face consultations in significantly more cases than by teleconsultations (P = 0.001). Where both types of consultation resulted in a single diagnosis there was a high level of agreement (kappa = 0.96, lower 95% confidence limit 0.91-1.00). Overall follow-up rates from both types of consultation were almost identical. Fifty per cent of patients seen could have been managed using a single videoconferenced teleconsultation without any requirement for further specialist intervention. Patients reported high levels of satisfaction with the teleconsultations. General practitioners reported that 75% of the teleconsultations were of educational benefit. This study illustrates the potential of telemedicine to diagnose and manage dermatology cases referred from primary care. Once the problem of image quality has been addressed, further studies will be required to investigate the cost-effectiveness of a teledermatology service and the potential consequences for the provision of dermatological services in the U.K.

A connectionist perspective of the transition from Face-to-Face to online teaching in Higher Education

Existing research shows a slow transition to online education by many university teaching staff. A mixed methods approach is used to survey teacher educators in three jurisdictions in the UK who have made the transition to online teaching, followed by focus group and individual interviews to triangulate the data. The eight tenets of connectivism are used as a lens for analysis. Findings reveal sound pedagogical reasons for the limited choice of online tools and tutors highlight two elements, namely, self-fulfilment and their desire to continually develop as an educator, as the rationale for adopting informal professional development in the 21st century.

Classification of fundus autofluorescence patterns in early age-related macular disease

PURPOSE. To describe and classify patterns of abnormal fundus autofluorescence (FAF) in eyes with early nonexudative age-related macular disease (AMD). METHODS. FAF images were recorded in eyes with early AMD by confocal scanning laser ophthalmoscopy (cSLO) with excitation at 488 nm (argon or OPSL laser) and emission above 500 or 521 nm (barrier filter). A standardized protocol for image acquisition and generation of mean images after automated alignment was applied, and routine fundus photographs were obtained. FAF images were classified by two independent observers. The ? statistic was applied to assess intra- and interobserver variability. RESULTS. Alterations in FAF were classified into eight phenotypic patterns including normal, minimal change, focal increased, patchy, linear, lacelike, reticular, and speckled. Areas with abnormal increased or decreased FAF signals may or may not have corresponded to funduscopically visible alterations. For intraobserver variability, ? of observer I was 0.80 (95% confidence interval [CI]0.71-0.89) and of observer II, 0.74. (95% CI, 0.64-0.84). For interobserver variability, ? was 0.77 (95% CI, 0.67-0.87). CONCLUSIONS. Various phenotypic patterns of abnormal FAF can be identified with cSLO imaging. Distinct patterns may reflect heterogeneity at a cellular and molecular level in contrast to a nonspecific aging process. The results indicate that the classification system yields a relatively high degree of intra- and interobserver agreement. It may be applicable for determination of novel prognostic determinants in longitudinal natural history studies, for identification of genetic risk factors, and for monitoring of future therapeutic interventions to slow the progression of early AMD. Copyright © Association for Research in Vision and Ophthalmology.

Iris cysts in children:Classification, incidence, and management. The 1998 Torrence A Makley Jr Lecture

Background - Iris cysts in children are uncommon and there is relatively little information on their classification, incidence, and management. Methods - The records of all children under age 20 years who were diagnosed with iris cyst were reviewed and the types and incidence of iris cysts of childhood determined. Based on these observations recommendations were made regarding management of iris cysts in children. Results - Of 57 iris cysts in children, 53 were primary and four were secondary. There were 44 primary cysts of the iris pigment epithelium, 34 of which were of the peripheral or iridociliary type, accounting for 59% of all childhood iris cysts. It was most commonly diagnosed in the teenage years, more common in girls (68%), was not recognised in infancy, remained stationary or regressed, and required no treatment. The five mid-zonal pigment epithelial cysts were diagnosed at a mean age of 14 years, were more common in boys (83%), remained stationary, and required no treatment. The pupillary type of pigment epithelial cyst was generally recognised in infancy and, despite involvement of the pupillary aperture, also required no treatment. There were nine cases of primary iris stromal cysts, accounting for 16% of all childhood iris cysts. This cyst was usually diagnosed in infancy, was generally progressive, and required treatment in eight of the nine cases, usually by aspiration and cryotherapy or surgical resection. Among the secondary iris cysts, two were post-traumatic epithelial ingrowth cysts and two were tumour induced cysts, one arising from an intraocular lacrimal gland choristoma and one adjacent to a peripheral iris naevus. Conclusions - Most iris cysts of childhood are primary pigment epithelial cysts and require no treatment. However, the iris stromal cyst, usually recognised in infancy, is generally an aggressive lesion that requires treatment by aspiration or surgical excision.

Le survivant de la Shoah face au texte de fiction: un écran protecteur ou un écran projecteur? L’exemple d’Anna Langfus

La terminologie ‘écriture écran’ est souvent utilisée dans un sens proche de celui que lui donne Annie Ernaux lorsqu’elle écrit que ‘la fiction protège’ en permettant à un auteur de dire tout en gardant le lecteur à distance. Pourtant, de Blanchot à Genette, de nombreux critiques ont souligné que le texte est par essence un espace qui n’existe que dans et par cet échange, le lecteur – surtout dans le cas des textes de fiction – devant s’investir, se projeter dans le texte lu. Le texte de fiction serait-il donc un écran protecteur pour celui qui tient la plume et un écran projecteur pour celui qui tient le livre ? En nous basant principalement sur des textes de la psychanalyste Rachel Rosenblum et de l’auteure et survivante de la Shoah Anna Langfus, nous suggèrerons que, pour l’auteur comme pour le lecteur, le texte de fiction est à la fois un écran protecteur et un écran projecteur, ces deux fonctions étant étroitement liées et nullement contradictoires. Nous montrerons en effet qu’aucun genre n’est a priori protecteur puisque c’est l’acte de lecture ou d’écriture qui peut se transformer en morbide compulsion de répétition quand la mémoire d’un lecteur ou d’un auteur est devenue pathologique.

Protected area networks and savannah bird biodiversity in the face of climate change and land degradation

The extent to which climate change might diminish the efficacy of protected areas is one of the most pressing conservation questions. Many projections suggest that climate-driven species distribution shifts will leave protected areas impoverished and species inadequately protected while other evidence suggests that intact ecosystems within protected areas will be resilient to change. Here, we tackle this problem empirically. We show how recent changes in distribution of 139 Tanzanian savannah bird species are linked to climate change, protected area status and land degradation. We provide the first evidence of climate-driven range shifts for an African bird community. Our results suggest that the continued maintenance of existing protected areas is an appropriate conservation response to the challenge of climate and environmental change.

Erythrocytosis in children and adolescents-classification, characterization, and consensus recommendations for the diagnostic approach

During recent years, the increasing knowledge of genetic and physiological changes in polycythemia vera (PV) and of different types of congenital erythrocytosis has led to fundamental changes in recommendations for the diagnostic approach to patients with erythrocytosis. Although widely accepted for adult patients this approach may not be appropriate with regard to children and adolescents affected by erythrocytosis. The "congenital erythrocytosis" working group established within the framework of the MPN&MPNr-EuroNet (COST action BM0902) addressed this question in a consensus finding process and developed a specific algorithm for the diagnosis of erythrocytosis in childhood and adolescence which is presented here. Pediatr Blood Cancer 2013;9999:XX-XX. © 2013 Wiley Periodicals, Inc.

Analysis of Bayesian classification-based approaches for Android malware detection

Mobile malware has been growing in scale and complexity spurred by the unabated uptake of smartphones worldwide. Android is fast becoming the most popular mobile platform resulting in sharp increase in malware targeting the platform. Additionally, Android malware is evolving rapidly to evade detection by traditional signature-based scanning. Despite current detection measures in place, timely discovery of new malware is still a critical issue. This calls for novel approaches to mitigate the growing threat of zero-day Android malware. Hence, the authors develop and analyse proactive machine-learning approaches based on Bayesian classification aimed at uncovering unknown Android malware via static analysis. The study, which is based on a large malware sample set of majority of the existing families, demonstrates detection capabilities with high accuracy. Empirical results and comparative analysis are presented offering useful insight towards development of effective static-analytic Bayesian classification-based solutions for detecting unknown Android malware.