<p>Face detection is a preprocessing step for face recognition algorithms. It is the localization of face/faces in an image or image sequence. Once the face(s) are localized, other computer vision algorithms such as face recognition, image compression, camera auto focusing etc are </p><p>applied. Because of the multiple usage areas, there are many research efforts in face processing. Face detection is a challenging computer vision problem because of lighting conditions, a high degree of variability in size, shape, background, color, etc. To build fully </p><p>automated systems, robust and efficient face detection algorithms are required. </p><p> </p><p>Numerous techniques have been developed to detect faces in a single image; in this project we have used a classification-based face detection method using Gabor filter features. We have designed five frequencies corresponding to eight orientations channels for extracting facial features from local images. The feature vector based on Gabor filter is used as the input of the face/non-face classifier, which is a Support Vector Machine (SVM) on a reduced feature </p><p>subspace extracted by using principal component analysis (PCA). </p><p> </p><p>Experimental results show promising performance especially on single face images where 78% accuracy is achieved with 0 false acceptances.</p>

Face detection is a preprocessing step for face recognition algorithms. It is the localization of face/faces in an image or image sequence. Once the face(s) are localized, other computer vision algorithms such as face recognition, image compression, camera auto focusing etc are applied. Because of the multiple usage areas, there are many research efforts in face processing. Face detection is a challenging computer vision problem because of lighting conditions, a high degree of variability in size, shape, background, color, etc. To build fully automated systems, robust and efficient face detection algorithms are required. Numerous techniques have been developed to detect faces in a single image; in this project we have used a classification-based face detection method using Gabor filter features. We have designed five frequencies corresponding to eight orientations channels for extracting facial features from local images. The feature vector based on Gabor filter is used as the input of the face/non-face classifier, which is a Support Vector Machine (SVM) on a reduced feature subspace extracted by using principal component analysis (PCA). Experimental results show promising performance especially on single face images where 78% accuracy is achieved with 0 false acceptances.

No description available.

Stages in Drosophila development denote the time after fertilization at which certain specific events occur in the developmental cycle. Stage information of a host embryo, as well as spatial information of a gene expression region is indispensable input for the discovery of the pattern of gene-gene interaction. Manual labeling of stages is becoming a bottleneck under the circumstance of high throughput embryo images. Automatic recognition based on the appearances of embryos is becoming a more desirable scheme. This problem, however, is very challenging due to severe variations of illumination and gene expressions. In this research thesis, we propose an appearance based recognition method using orientation histograms and Gabor filter. Furthermore, we apply Principal Component Analysis to reduce the dimension of the low-level features, aiming to accelerate the speed of recognition. With the experiments on BDGP images, we show the promise of the proposed method.

In this thesis work, a detailed analysis of biometric technologies has been done and a new palmprint recognition algorithm has been implemented. The proposed algorithm is based on 2-D Gabor filters. The developed algorithm is first tested on The Hong Kong Polytechnic University Palmprint Database in terms of accuracy, speed and template size. Then a scanner is integrated into the developed algorithm in order to acquire palm images / in this way an online palmprint recognition system has been developed. Then a small palmprint database is formed via this system in Middle East Technical University. Results on this new database have also shown the success of the developed algorithm.

Orientador: Roberto de Alencar Lotufo / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Eletrica e de Computação / Made available in DSpace on 2018-08-07T11:27:50Z (GMT). No. of bitstreams: 1 Sa_GustavoFerreiraCardosode_M.pdf: 2313623 bytes, checksum: ca6abbf3a186c9d5bed2d6b1e73e3a9f (MD5) Previous issue date: 2006 / Resumo: Neste trabalho são apresentadas melhorias na robustez do método FingerCode para reconhecimento de impressões digitais. No FingerCode a textura dos componentes orientados das impressões digitais são extraídas por um banco direcional de filtros Gabor. Posteriormente, os componentes orientados são setorizados e para cada setor é computado um valor. Este conjunto de valores forma o vetor de atributos. Finalmente, a média da diferença absoluta dos dois vetores de atributos é computada indicando a similaridade entre duas impressões digitais. Foram testadas várias soluções e entre as que apresentaram melhores resultados destacam-se: a substituição dos valores dos atributos através de uma função não-linear, a ponderação dos valores de atributo de acordo com características estatísticas da distribuição espacial dos valores e o cálculo de medidas estatísticas extraídas dos histogramas de distribuição de diferenças. Estas funções apresentaram um ganho significativo, principalmente para o caso dos sensores óticos com uma melhoria de aproximadamente 45% no EER. Outra contribuição apresentada foi uma nova implementação rápida do filtro Gabor 2D, que se constitui de uma onda sinusoidal modulada por um envelope gaussiano. A filtragem 2D da imagem por um banco de filtros Gabor 2D é uma das etapas de maior consumo de tempo no processamento de imagens. Na nova solução proposta, o filtro Gabor 2D é separado em dois filtros Gabor 1D ortogonais, bastando para isto que o envelope gaussiano obedeça a condição de ser circular. O processamento com o filtro separado é mais rápida do que com o filtro não-separado e o ganho na performance aumenta à medida que aumenta o tamanho da imagem ou do filtro. Também foram desenvolvidas novas técnicas de segmentação: baseada em morfologia matemática e baseada em filtros Gabor. Estas segmentações ocorrem ao nível do píxel, com ótimos resultados, principalmente após a uniformização da área através de processos morfológicos / Abstract: In this work it is introduced improvements in robustness of FingerCode method to recognize fingerprints. In the FingerCode the texture of fingerprint oriented components are extracted by a bank of directional Gabor filters. After that, the oriented components are tessellated and a value is computed for each sector. This set of values constitutes the attribute vector. Finally, the absolute difference mean between the two attribute vectors is computed that gives the similarity between two fingerprints. New solutions were tested; among them the best results were obtained by: attribute values replacement by a non-linear function, attribute values weighting by statistical characteristics of spatial distribution of values, and the calculus of statistical measures extracted from the difference distribution histograms. These functions presented a significant gain, mainly in the case of optical sensors with an improvement about 45% in EER. Another contribution presented was a new fast implementation of the 2D Gabor filter, which constitutes in a sinusoidal wave modulated by a Gaussian envelope. The 2D image filtering by a bank of 2D Gabor filters is one of the most expensive stage of image processing. In the new solution proposed, the 2D Gabor filter is separated in two orthogonal 1D Gabor filters, for this the Gaussian envelope must obey the condition of being circular. Processing with the separated filter is faster than the non-separated filter, and the gain improves as the size of image or filter increases. Also it was developed new segmentation techniques: based on mathematical morphology, and based on Gabor filters. Those segmentations occur at pixel level, with good results, mostly after the area regularization with morphological processes / Mestrado / Engenharia de Computação / Mestre em Engenharia Elétrica

No description available.

O câncer de mama é a doença que mais acomete as mulheres em todo o mundo, sendo o tratamento mais eficaz se for diagnosticada em estágio inicial. A partir de 2011, nos programas de rastreamento de países desenvolvidos, vem sendo empregada uma nova modalidade de exame, a tomossíntese digital mamária (Digital Breast Tomosynthesis - DBT), que possui diversas vantagens se comparada à mamografia digital. No exame, o médico radiologista busca por sinais suspeitos na imagem, como: nódulos, microcalcificações e distorção arquitetural mamária (DAM). Sendo que, este último pode representar o estágio mais inicial de um câncer em formação, podendo se manifestar antes da formação de qualquer outra lesão. No entanto, a DAM é difícil de ser detectada pois modifica o tecido mamário de forma sutil, não havendo qualquer formação de massa ou a borda definida. Os sistemas computacionais de auxílio ao diagnóstico (Computer-Aided Detection - CAD) vêm apresentando alto desempenho na detecção de nódulos e microcalcificações mamárias, mas para o caso da DAM, o desempenho ainda é insatisfatório. Algumas limitações são normalmente reportadas nos algoritmos adotados para detectar automaticamente a DAM. O presente trabalho tem por objetivo propor novas abordagens para aumentar a precisão dos métodos computacionais de detecção: o uso de descritores de micro-padrões local para discriminação de áreas suspeitas; redução de falsos-positivos; uso do volume 3D fornecido pelo exame de DBT e; uso de arquitetura de aprendizagem profunda para discriminação e classificação de regiões suspeitas. Os diversos testes efetuados em cada proposta mostraram que é possível melhorar as taxas de detecção da DAM, mesmo para imagens de DBT onde ainda não há um esquema computacional de detecção bem estabelecido. / Breast cancer is the disease that most affects women worldwide and is the most effective treatment if it is diagnosed at early stages. Since 2011, in developed countries screening programs has been employed a new exam, the digital breast tomosynthesis (DBT), which has several advantages compared to the digital mammography. In the exam, the radiologist looks for suspicious signs in the image such as masses, microcalcifications and architectural distortion of breast (ADB). Since the later may represent the earliest sign of a cancer in formation, it may manifests before the formation of any other lesion. However, ADB is difficult to be detected due to its subtly changes the breast tissue, with no mass or defined shape. Computer-aided detection (CAD) systems have shown good results in the detection of masses and microcalcifications, however, for ADB the performance is still poor. Several weakness are reported in the pipeline adopted to automatic detection of ADB. The present work aims to propose new approaches to increase the accuracy of the current CAD pipeline: the use of local micro-pattern descriptors to discriminate suspicious areas; false-positives reduction; automatic detection of ADB in DBT images using and tridimensionality of the exam and; use of deep learning architecture to discriminate and classify suspicious regions. The several tests performed on each proposal showed that it is possible to improve the detection rates even for DBT images where there is no established CAD pipeline.

This master thesis is focused on parametrization of image point neighborhood. Some methods for point localization and point descriptors are described and summarized. Gabor filter is described in detail. The practical part of thesis is chiefly concerned with particle filter tracking system. The weight of each particle is determined by the Gabor filter.

Object recognition is one of many tasks in which the computer is still behind the human. Therefore, development in this area takes inspiration from nature and especially from the function of the human brain. This work focuses on object recognition based on extracting relevant information from images, features. Features are obtained in a similar way as the human brain processes visual stimuli. Subsequently, these features are used to train classifiers for object recognition (e.g. SVM, k-NN, ANN). This work examines the feature extraction stage. Its aim is to improve the feature extraction and thereby increase performance of object recognition by computer.