A Technique for Digital Color Image Watermarking Using ICA

With the increase of use of digital media the need for the methods of multimedia protection becomes extremely important. The number of the solutions to the problem from encryption to watermarking is large and is growing every year. In this work digital image watermarking is considered, specifically a novel method of digital watermarking of color and spectral images. An overview of existing methods watermarking of color and grayscale images is given in the paper. Methods using independent component analysis (ICA) for detection and the ones using discrete wavelet transform (DWT) and discrete cosine transform (DCT) are considered in more detail. A novel method of watermarking proposed in this paper allows embedding of a color or spectral watermark image into color or spectral image consequently and successful extraction of the watermark out of the resultant watermarked image. A number of experiments have been performed on the quality of extraction depending on the parameters of the embedding procedure. Another set of experiments included the test of the robustness of the algorithm proposed. Three techniques have been chosen for that purpose: median filter, low-pass filter (LPF) and discrete cosine transform (DCT), which are a part of a widely known StirMark - Image Watermarking Robustness Test. The study shows that the proposed watermarking technique is fragile, i.e. watermark is altered by simple image processing operations. Moreover, we have found that the contents of the image to be watermarked do not affect the quality of the extraction. Mixing coefficients, that determine the amount of the key and watermark image in the result, should not exceed 1% of the original. The algorithm proposed has proven to be successful in the task of watermark embedding and extraction.

Characterization of fiber and vessel elements in pulp suspension images

The thesis is related to the topic of image-based characterization of fibers in pulp suspension during the papermaking process. Papermaking industry is focusing on process control optimization and automatization, which makes it possible to manufacture highquality products in a resource-efficient way. Being a part of the process control, pulp suspension analysis allows to predict and modify properties of the end product. This work is a part of the tree species identification task and focuses on analysis of fiber parameters in the pulp suspension at the wet stage of paper production. The existing machine vision methods for pulp characterization were investigated, and a method exploiting direction sensitive filtering, non-maximum suppression, hysteresis thresholding, tensor voting, and curve extraction from tensor maps was developed. Application of the method to the microscopic grayscale pulp images made it possible to detect curves corresponding to fibers in the pulp image and to compute their morphological characteristics. Performance of the method was evaluated based on the manually produced ground truth data. An accuracy of fiber characteristics estimation, including length, width, and curvature, for the acacia pulp images was found to be 84, 85, and 60% correspondingly.

Developing raster bands computed from LiDAR to improve forest

This study examines the use of di erent features derived from remotely sensed data in segmentation of forest stands. Surface interpolation methods were applied to LiDAR points in order to represent data in the form of grayscale images. Median and mean shift ltering was applied to the data for noise reduction. The ability of di erent compositions of rasters obtained from LiDAR data and an aerial image to maximize stand homogeneity in the segmentation was evaluated. The quality of forest stand delineations was assessed by the Akaike information criterion. The research was performed in co-operation with Arbonaut Ltd., Joensuu, Finland.

Deep generative models for facial keypoints detection

A new area of machine learning research called deep learning, has moved machine learning closer to one of its original goals: artificial intelligence and general learning algorithm. The key idea is to pretrain models in completely unsupervised way and finally they can be fine-tuned for the task at hand using supervised learning. In this thesis, a general introduction to deep learning models and algorithms are given and these methods are applied to facial keypoints detection. The task is to predict the positions of 15 keypoints on grayscale face images. Each predicted keypoint is specified by an (x,y) real-valued pair in the space of pixel indices. In experiments, we pretrained deep belief networks (DBN) and finally performed a discriminative fine-tuning. We varied the depth and size of an architecture. We tested both deterministic and sampled hidden activations and the effect of additional unlabeled data on pretraining. The experimental results show that our model provides better results than publicly available benchmarks for the dataset.

Spectral retinal image processing and analysis for ophthalmology

Diabetic retinopathy, age-related macular degeneration and glaucoma are the leading causes of blindness worldwide. Automatic methods for diagnosis exist, but their performance is limited by the quality of the data. Spectral retinal images provide a significantly better representation of the colour information than common grayscale or red-green-blue retinal imaging, having the potential to improve the performance of automatic diagnosis methods. This work studies the image processing techniques required for composing spectral retinal images with accurate reflection spectra, including wavelength channel image registration, spectral and spatial calibration, illumination correction, and the estimation of depth information from image disparities. The composition of a spectral retinal image database of patients with diabetic retinopathy is described. The database includes gold standards for a number of pathologies and retinal structures, marked by two expert ophthalmologists. The diagnostic applications of the reflectance spectra are studied using supervised classifiers for lesion detection. In addition, inversion of a model of light transport is used to estimate histological parameters from the reflectance spectra. Experimental results suggest that the methods for composing, calibrating and postprocessing spectral images presented in this work can be used to improve the quality of the spectral data. The experiments on the direct and indirect use of the data show the diagnostic potential of spectral retinal data over standard retinal images. The use of spectral data could improve automatic and semi-automated diagnostics for the screening of retinal diseases, for the quantitative detection of retinal changes for follow-up, clinically relevant end-points for clinical studies and development of new therapeutic modalities.

Information utile à la catégorisation du sexe des visages

La présente thèse avait pour mandat d’examiner la question suivante : quels sont les indices visuels utilisés pour catégoriser le sexe d’un visage et comment sont-ils traités par le cerveau humain? La plupart des études examinant l’importance de certaines régions faciales pour la catégorisation du sexe des visages présentaient des limites quant à leur validité externe. L’article 1 visait à investiguer l’utilisation des indices achromatiques et chromatiques (sur l’axe xy) dans un contexte de plus grande validité externe. Pour ce faire, nous avons utilisé la technique Bubbles afin d’échantillonner l’espace xy de visages en couleurs n’ayant subi aucune transformation. Afin d’éviter les problèmes liés à la grande répétition des mêmes visages, nous avons utilisé un grand nombre de visages (c.-à-d. 300 visages caucasiens d’hommes et de femmes) et chaque visage n’a été présenté qu’une seule fois à chacun des 30 participants. Les résultats indiquent que la région des yeux et des sourcils—probablement dans le canal blanc-noir—est l’indice le plus important pour discriminer correctement le genre des visages; et que la région de la bouche—probablement dans le canal rouge-vert—est l’indice le plus important pour discriminer rapidement et correctement le genre des visages. Plusieurs études suggèrent qu’un indice facial que nous n’avons pas étudié dans l’article 1—les distances interattributs—est crucial à la catégorisation du sexe. L’étude de Taschereau et al. (2010) présente toutefois des données allant à l’encontre de cette hypothèse : les performances d’identification des visages étaient beaucoup plus faibles lorsque seules les distances interattributs réalistes étaient disponibles que lorsque toutes les autres informations faciales à l’exception des distances interattributs réalistes étaient disponibles. Quoi qu’il en soit, il est possible que la faible performance observée dans la condition où seules les distances interattributs étaient disponibles soit explicable non par une incapacité d’utiliser ces indices efficacement, mais plutôt par le peu d’information contenue dans ces indices. L’article 2 avait donc comme objectif principal d’évaluer l’efficacité—une mesure de performance qui compense pour la faiblesse de l’information disponible—des distances interattributs réalistes pour la catégorisation du sexe des visages chez 60 participants. Afin de maximiser la validité externe, les distances interattributs manipulées respectaient la distribution et la matrice de covariance observées dans un large échantillon de visages (N=515). Les résultats indiquent que les efficacités associées aux visages ne possédant que de l’information au niveau des distances interattributs sont un ordre de magnitude plus faibles que celles associées aux visages possédant toute l’information que possèdent normalement les visages sauf les distances interattributs et donnent le coup de grâce à l’hypothèse selon laquelle les distances interattributs seraient cuciale à la discrimination du sexe des visages. L’article 3 avait pour objectif principal de tester l’hypothèse formulée à la fin de l’article 1 suivant laquelle l’information chromatique dans la région de la bouche serait extraite très rapidement par le système visuel lors de la discrimination du sexe. Cent douze participants ont chacun complété 900 essais d’une tâche de discrimination du genre pendant laquelle l’information achromatique et chromatique des visages était échantillonnée spatiotemporellement avec la technique Bubbles. Les résultats d’une analyse présentée en Discussion seulement confirme l’utilisation rapide de l’information chromatique dans la région de la bouche. De plus, l’utilisation d’un échantillonnage spatiotemporel nous a permis de faire des analyses temps-fréquences desquelles a découlé une découverte intéressante quant aux mécanismes d’encodage des informations spatiales dans le temps. Il semblerait que l’information achromatique et chromatique à l’intérieur d’une même région faciale est échantillonnée à la même fréquence par le cerveau alors que les différentes parties du visage sont échantillonnées à des fréquences différentes (entre 6 et 10 Hz). Ce code fréquentiel est compatible avec certaines évidences électrophysiologiques récentes qui suggèrent que les parties de visages sont « multiplexées » par la fréquence d’oscillations transitoires synchronisées dans le cerveau.

Role of color in face recognition

One of the key challenges in face perception lies in determining the contribution of different cues to face identification. In this study, we focus on the role of color cues. Although color appears to be a salient attribute of faces, past research has suggested that it confers little recognition advantage for identifying people. Here we report experimental results suggesting that color cues do play a role in face recognition and their contribution becomes evident when shape cues are degraded. Under such conditions, recognition performance with color images is significantly better than that with grayscale images. Our experimental results also indicate that the contribution of color may lie not so much in providing diagnostic cues to identity as in aiding low-level image-analysis processes such as segmentation.

A low cost one-camera optical tracking system for indoor wide-area augmented and virtual reality environments

In the last years the number of industrial applications for Augmented Reality (AR) and Virtual Reality (VR) environments has significantly increased. Optical tracking systems are an important component of AR/VR environments. In this work, a low cost optical tracking system with adequate attributes for professional use is proposed. The system works in infrared spectral region to reduce optical noise. A highspeed camera, equipped with daylight blocking filter and infrared flash strobes, transfers uncompressed grayscale images to a regular PC, where image pre-processing software and the PTrack tracking algorithm recognize a set of retro-reflective markers and extract its 3D position and orientation. Included in this work is a comprehensive research on image pre-processing and tracking algorithms. A testbed was built to perform accuracy and precision tests. Results show that the system reaches accuracy and precision levels slightly worse than but still comparable to professional systems. Due to its modularity, the system can be expanded by using several one-camera tracking modules linked by a sensor fusion algorithm, in order to obtain a larger working range. A setup with two modules was built and tested, resulting in performance similar to the stand-alone configuration.

Chromatin and morphometry alterations of bovine sperm head after Percoll (TM) density gradient centrifugation

The aim of this work was to assess the morphometry and chromatin integrity of bovine sperm head after a three layers discontinuous Percoll (TM) density gradient centrifugation. Frozen semen doses were obtained from six bulls of different breeds, including three taurine and three Zebu animals. Three ejaculates per bull were evaluated. The semen doses were thawed and two smears were made from each sample before (control) and after the Percoll (TM) centrifugation (Percoll (TM) group). The smears were stained with toluidine blue and grayscale digital images were captured and processed in Scilab environment software. It was observed that chromatin heterogeneity was reduced (P<0.05) and chromatin decondensation was increased (P<0.05) after the Percollm treatment utilized. In addition, it was observed that sperm head length was higher (P<0.05) and the side symmetry was lower (P<0.05) in centrifuged sperm cells. When analyzed separately by subspecies, it was observed that the decrease (P<0.05) in chromatin heterogeneity and the increase (P<0.05) in chromatin decondensation occurred in Zebu sperm heads. In addition, the length and the width:length ratio of sperm heads was affected by Percoll (TM) centrifugation in Zebu semen. In conclusion, the three layers discontinuous Percoll (TM) centrifugation increased the chromatin decondensation and the morphometric alterations of frozen-thawed bovine semen. However, the real implication of these findings in fertility rates of centrifuged sperm must be investigated. (C) 2011 Elsevier B.V. All rights reserved.

Extended depth from focus reconstruction using NIH ImageJ plugins: Quality and resolution of elevation maps

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

Algoritmos de detecção de bordas implementados em FPGA

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)

Eliminação de ruído impulsivo em imagens coloridas usando um filtro mediano seletivo e retoque digital

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

Ultra-sonografia na avaliação andrológica de bovinos da raça Nelore

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)

ARFI elastography as a complementary diagnostic method for mammary neoplasia in female dogs - preliminary results

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

Effect of digitization parameters on periapical radiographic image quality with regard to anatomic landmarks

The aim of this study was to evaluate the influence of digitization parameters on periapical radiographic image quality, with regard to anatomic landmarks. Digitized images (n = 160) were obtained using a flatbed scanner with resolutions of 300, 600 and 2400 dpi. The radiographs of 2400 dpi were decreased to 300 and 600 dpi before storage. Digitizations were performed with and without black masking using 8-bit and 16-bit grayscale and saved in TIFF format. Four anatomic landmarks were classified by two observers (very good, good, moderate, regular, poor), in two random sessions. Intraobserver and interobserver agreements were evaluated by Kappa statistics. Inter and intraobserver agreements ranged according to the anatomic landmarks and resolution used. The results obtained demonstrated that the cement enamel junction was the anatomic landmark that presented the poorest concordance. The use of black masking provided better results in the digitized image. The use of a mask to cover radiographs during digitization is necessary. Therefore, the concordance ranged from regular to moderate for the intraobserver evaluation and concordance ranged from regular to poor for interobserver evaluation.