Détection de mouvement par modèle biologique de fusion de donnée inspiré de la rétine humaine

Ce mémoire s'intéresse à la détection de mouvement dans une séquence d'images acquises à l'aide d'une caméra fixe. Dans ce problème, la difficulté vient du fait que les mouvements récurrents ou non significatifs de la scène tels que les oscillations d'une branche, l'ombre d'un objet ou les remous d'une surface d'eau doivent être ignorés et classés comme appartenant aux régions statiques de la scène. La plupart des méthodes de détection de mouvement utilisées à ce jour reposent en fait sur le principe bas-niveau de la modélisation puis la soustraction de l'arrière-plan. Ces méthodes sont simples et rapides mais aussi limitées dans les cas où l'arrière-plan est complexe ou bruité (neige, pluie, ombres, etc.). Cette recherche consiste à proposer une technique d'amélioration de ces algorithmes dont l'idée principale est d'exploiter et mimer deux caractéristiques essentielles du système de vision humain. Pour assurer une vision nette de l’objet (qu’il soit fixe ou mobile) puis l'analyser et l'identifier, l'œil ne parcourt pas la scène de façon continue, mais opère par une série de ``balayages'' ou de saccades autour (des points caractéristiques) de l'objet en question. Pour chaque fixation pendant laquelle l'œil reste relativement immobile, l'image est projetée au niveau de la rétine puis interprétée en coordonnées log polaires dont le centre est l'endroit fixé par l'oeil. Les traitements bas-niveau de détection de mouvement doivent donc s'opérer sur cette image transformée qui est centrée pour un point (de vue) particulier de la scène. L'étape suivante (intégration trans-saccadique du Système Visuel Humain (SVH)) consiste ensuite à combiner ces détections de mouvement obtenues pour les différents centres de cette transformée pour fusionner les différentes interprétations visuelles obtenues selon ses différents points de vue.

classification and content based retrieval of digital mammograms and placental sonograms

Image processing has been a challenging and multidisciplinary research area since decades with continuing improvements in its various branches especially Medical Imaging. The healthcare industry was very much benefited with the advances in Image Processing techniques for the efficient management of large volumes of clinical data. The popularity and growth of Image Processing field attracts researchers from many disciplines including Computer Science and Medical Science due to its applicability to the real world. In the meantime, Computer Science is becoming an important driving force for the further development of Medical Sciences. The objective of this study is to make use of the basic concepts in Medical Image Processing and develop methods and tools for clinicians’ assistance. This work is motivated from clinical applications of digital mammograms and placental sonograms, and uses real medical images for proposing a method intended to assist radiologists in the diagnostic process. The study consists of two domains of Pattern recognition, Classification and Content Based Retrieval. Mammogram images of breast cancer patients and placental images are used for this study. Cancer is a disaster to human race. The accuracy in characterizing images using simplified user friendly Computer Aided Diagnosis techniques helps radiologists in detecting cancers at an early stage. Breast cancer which accounts for the major cause of cancer death in women can be fully cured if detected at an early stage. Studies relating to placental characteristics and abnormalities are important in foetal monitoring. The diagnostic variability in sonographic examination of placenta can be overlooked by detailed placental texture analysis by focusing on placental grading. The work aims on early breast cancer detection and placental maturity analysis. This dissertation is a stepping stone in combing various application domains of healthcare and technology.

Changes of traditional farming systems and their effects on land degradation and socio-economic conditions in the Inle Lake region, Myanmar

At many locations in Myanmar, ongoing changes in land use have negative environmental impacts and threaten natural ecosystems at local, regional and national scales. In particular, the watershed area of Inle Lake in eastern Myanmar is strongly affected by the environmental effects of deforestation and soil erosion caused by agricultural intensification and expansion of agricultural land, which are exacerbated by the increasing population pressure and the growing number of tourists. This thesis, therefore, focuses on land use changes in traditional farming systems and their effects on socio-economic and biophysical factors to improve our understanding of sustainable natural resource management of this wetland ecosystem. The main objectives of this research were to: (1) assess the noticeable land transformations in space and time, (2) identify the typical farming systems as well as the divergent livelihood strategies, and finally, (3) estimate soil erosion risk in the different agro-ecological zones surrounding the Inle Lake watershed area. GIS and remote sensing techniques allowed to identify the dynamic land use and land cover changes (LUCC) during the past 40 years based on historical Corona images (1968) and Landsat images (1989, 2000 and 2009). In this study, 12 land cover classes were identified and a supervised classification was used for the Landsat datasets, whereas a visual interpretation approach was conducted for the Corona images. Within the past 40 years, the main landscape transformation processes were deforestation (- 49%), urbanization (+ 203%), agricultural expansion (+ 34%) with a notably increase of floating gardens (+ 390%), land abandonment (+ 167%), and marshlands losses in wetland area (- 83%) and water bodies (- 16%). The main driving forces of LUCC appeared to be high population growth, urbanization and settlements, a lack of sustainable land use and environmental management policies, wide-spread rural poverty, an open market economy and changes in market prices and access. To identify the diverse livelihood strategies in the Inle Lake watershed area and the diversity of income generating activities, household surveys were conducted (total: 301 households) using a stratified random sampling design in three different agro-ecological zones: floating gardens (FG), lowland cultivation (LL) and upland cultivation (UP). A cluster and discriminant analysis revealed that livelihood strategies and socio-economic situations of local communities differed significantly in the different zones. For all three zones, different livelihood strategies were identified which differed mainly in the amount of on-farm and off-farm income, and the level of income diversification. The gross margin for each household from agricultural production in the floating garden, lowland and upland cultivation was US$ 2108, 892 and 619 ha-1 respectively. Among the typical farming systems in these zones, tomato (Lycopersicon esculentum L.) plantation in the floating gardens yielded the highest net benefits, but caused negative environmental impacts given the overuse of inorganic fertilizers and pesticides. The Revised Universal Soil Loss Equation (RUSLE) and spatial analysis within GIS were applied to estimate soil erosion risk in the different agricultural zones and for the main cropping systems of the study region. The results revealed that the average soil losses in year 1989, 2000 and 2009 amounted to 20, 10 and 26 t ha-1, respectively and barren land along the steep slopes had the highest soil erosion risk with 85% of the total soil losses in the study area. Yearly fluctuations were mainly caused by changes in the amount of annual precipitation and the dynamics of LUCC such as deforestation and agriculture extension with inappropriate land use and unsustainable cropping systems. Among the typical cropping systems, upland rainfed rice (Oryza sativa L.) cultivation had the highest rate of soil erosion (20 t ha-1yr-1) followed by sebesten (Cordia dichotoma) and turmeric (Curcuma longa) plantation in the UP zone. This study indicated that the hotspot region of soil erosion risk were upland mountain areas, especially in the western part of the Inle lake. Soil conservation practices are thus urgently needed to control soil erosion and lake sedimentation and to conserve the wetland ecosystem. Most farmers have not yet implemented soil conservation measures to reduce soil erosion impacts such as land degradation, sedimentation and water pollution in Inle Lake, which is partly due to the low economic development and poverty in the region. Key challenges of agriculture in the hilly landscapes can be summarized as follows: fostering the sustainable land use of farming systems for the maintenance of ecosystem services and functions while improving the social and economic well-being of the population, integrated natural resources management policies and increasing the diversification of income opportunities to reduce pressure on forest and natural resources.

Geometry and Photometry in 3D Visual Recognition

The report addresses the problem of visual recognition under two sources of variability: geometric and photometric. The geometric deals with the relation between 3D objects and their views under orthographic and perspective projection. The photometric deals with the relation between 3D matte objects and their images under changing illumination conditions. Taken together, an alignment-based method is presented for recognizing objects viewed from arbitrary viewing positions and illuminated by arbitrary settings of light sources.

Improving Clustering Algorithms for Image Segmentation using Contour and Region Information

In image segmentation, clustering algorithms are very popular because they are intuitive and, some of them, easy to implement. For instance, the k-means is one of the most used in the literature, and many authors successfully compare their new proposal with the results achieved by the k-means. However, it is well known that clustering image segmentation has many problems. For instance, the number of regions of the image has to be known a priori, as well as different initial seed placement (initial clusters) could produce different segmentation results. Most of these algorithms could be slightly improved by considering the coordinates of the image as features in the clustering process (to take spatial region information into account). In this paper we propose a significant improvement of clustering algorithms for image segmentation. The method is qualitatively and quantitative evaluated over a set of synthetic and real images, and compared with classical clustering approaches. Results demonstrate the validity of this new approach

Colour image of the thorax and abdomen

This is an image taken from Anatomy tv, an interactive resource for teaching and learning in anatomy and physiology which the University Library subscribes to. This image may not be changed, but you may take a copy and present it with other materials and resources you are using so long as they are password protected for access by members of the University only. "All products and all images within the products are protected by copyright. The products and images can only be used for private educational purposes, unless a specific license is purchased for any other usage. For any commercial usage of the images, please contact Primal Pictures Limited. The products allow members of the University of Southampton to ‘copy and paste’ all of the text as well as the images in the 3D-model window and all of the slides. These can then be pasted into nearly any other word-processing or graphics program, including Powerpoint. These resources can be made available to members of the University of Southampton via a password-protected service. This again is designed solely as a service for private educational uses. Like any publisher, Primal Pictures protects itself against copyright infringement. Please do contact Debra Morris in the University Library before using these resources to ensure that conditions are respected. ©Primal Pictures Limited 2007

Land-cover classification from airborne LIDAR data fused with aerial optical images

Airborne LIght Detection And Ranging (LIDAR) provides accurate height information for objects on the earth, which makes LIDAR become more and more popular in terrain and land surveying. In particular, LIDAR data offer vital and significant features for land-cover classification which is an important task in many application domains. In this paper, an unsupervised approach based on an improved fuzzy Markov random field (FMRF) model is developed, by which the LIDAR data, its co-registered images acquired by optical sensors, i.e. aerial color image and near infrared image, and other derived features are fused effectively to improve the ability of the LIDAR system for the accurate land-cover classification. In the proposed FMRF model-based approach, the spatial contextual information is applied by modeling the image as a Markov random field (MRF), with which the fuzzy logic is introduced simultaneously to reduce the errors caused by the hard classification. Moreover, a Lagrange-Multiplier (LM) algorithm is employed to calculate a maximum A posteriori (MAP) estimate for the classification. The experimental results have proved that fusing the height data and optical images is particularly suited for the land-cover classification. The proposed approach works very well for the classification from airborne LIDAR data fused with its coregistered optical images and the average accuracy is improved to 88.9%.

Counting Her Dresses and other plays

Counting Her Dresses and other plays, devised by Lib Taylor, is a short, mixed-media, promenade performance based on theatre writings by Gertrude Stein. Gertrude Stein was an American writer who lived in Paris for most of her adult life and was a significant figure in the development of early twentieth century modernism. She was an artists’ patron and one of the first collectors of paintings by Picasso, Matisse, Cezanne and others. She wrote novels, short stories, poetry and criticism, as well as a number of very short ‘plays’ and opera libretti. Even though her plays are very rarely performed they have been very influential in the development of avant-garde theatre. The performance juxtaposes a soundcape of voices, live performance, an ‘exhibition’ of paintings and mediated images projected onto and across the space to evoke Stein’s sense of theatre as a place of experience and emotion, not as a place of story and action. The performance comprises five theatre fragments which have been combined in a collage that alludes to the experimental art that Stein promoted.

Watching the words: drama and poems

Some poems are inherently dramatic due to their narrative content or the events, characters, places and emotions that are their subject. Others have the potential for dramatisation because of some aural or visual quality of their poetic form. However, if dramatising poems is to be meaningful and effective children need to be taught something about the art form of drama rather than just being left to their own devices. This chapter explores the learning potential of considering the printed text of a poem as a notation of sound, movement, gesture and use of space. The chapter recognises a progression from simple nursery rhymes to the sophisticated use of poetic language in different types of literature that is mirrored in the journey from infants’ clapping games to the dramatic juxtaposition of aural and visual images in theatre and the performing arts.

Sustained striatal activity predicts eudaimonic well-being and cortisol output

Eudaimonic well-being—a sense of purpose, meaning, and engagement with life—is protective against psychopathology and predicts physical health, including lower levels of the stress hormone cortisol. Although it has been suggested that the ability to engage the neural circuitry of reward may promote well-being and mediate the relationship between well-being and health, this hypothesis has remained untested. To test this hypothesis, we had participants view positive, neutral, and negative images while fMRI data were collected. Individuals with sustained activity in the striatum and dorsolateral prefrontal cortex to positive stimuli over the course of the scan session reported greater well-being and had lower cortisol output. This suggests that sustained engagement of reward circuitry in response to positive events underlies well-being and adaptive regulation of the hypothalamic-pituitary-adrenal axis.

ReadingAct RGB-D action dataset and human action recognition from local features

For general home monitoring, a system should automatically interpret people’s actions. The system should be non-intrusive, and able to deal with a cluttered background, and loose clothes. An approach based on spatio-temporal local features and a Bag-of-Words (BoW) model is proposed for single-person action recognition from combined intensity and depth images. To restore the temporal structure lost in the traditional BoW method, a dynamic time alignment technique with temporal binning is applied in this work, which has not been previously implemented in the literature for human action recognition on depth imagery. A novel human action dataset with depth data has been created using two Microsoft Kinect sensors. The ReadingAct dataset contains 20 subjects and 19 actions for a total of 2340 videos. To investigate the effect of using depth images and the proposed method, testing was conducted on three depth datasets, and the proposed method was compared to traditional Bag-of-Words methods. Results showed that the proposed method improves recognition accuracy when adding depth to the conventional intensity data, and has advantages when dealing with long actions.

INFLUENCE OF INTERIMPLANT DISTANCES AND PLACEMENT DEPTH ON PAPILLA FORMATION AND CRESTAL RESORPTION: A CLINICAL AND RADIOGRAPHIC STUDY IN DOGS

Among the factors that contribute to the papilla formation and crestal bone preservation between contiguous implants, this animal study clinically and radiographically evaluated the interimplant distances (IDs) of 2 and 3 mm and the placement depths of Morse cone connection implants restored with platform switch. Bilateral mandibular premolars of 6 dogs were extracted, and after 12 weeks, the implants were placed. Four experimental groups were constituted: subcrestally with ID of 2 mm (2 SCL) and 3 mm (3 SCL) and crestally with ID of 2 mm (2 CL) and 3 mm (3 CL). Metallic crowns were immediately installed with a distance of 3 mm between the contact point and the bone crest. Eight weeks later, clinical measurements were performed to evaluate papilla formation, and radiographic images were taken to analyze the crestal bone remodeling. The subcrestal groups achieved better levels of papillae formation when compared with the crestal groups, with a significant difference between the 3 SCL and 3 CL groups (P = .026). Radiographically, the crestal bone preservation was also better in the subcrestal groups, with statistically significant differences between the 2SCL and 2CL groups (P = .002) and between the 3SCL and 3CL groups (P = .008). With the present conditions, it could be concluded that subcrestal implant placement had a positive impact on papilla formation and crestal bone preservation, which could favor the esthetic of anterior regions. However, the IDs of 2 and 3 mm did not show significantly different results.

HI aperture synthesis and optical observations of the pair of galaxies NGC 6907 and 6908

NGC 6908, an S0 galaxy situated in the direction of NGC 6907, was only recently recognized as a distinct galaxy, instead of only a part of NGC 6907. We present 21-cm radio synthesis observations obtained with the Giant Metrewave Radio Telescope (GMRT) and optical images and spectroscopy obtained with the Gemini-North telescope of this pair of interacting galaxies. From the radio observations, we obtained the velocity field and the H I column density map of the whole region containing the NGC 6907/8 pair, and by means of the Gemini multi-object spectroscopy we obtained high-quality photometric images and 5 angstrom resolution spectra sampling the two galaxies. By comparing the rotation curve of NGC 6907 obtained from the two opposite sides around the main kinematic axis, we were able to distinguish the normal rotational velocity field from the velocity components produced by the interaction between the two galaxies. Taking into account the rotational velocity of NGC 6907 and the velocity derived from the absorption lines for NGC 6908, we verified that the relative velocity between these systems is lower than 60 km s(-1). The emission lines observed in the direction of NGC 6908, not typical of S0 galaxies, have the same velocity expected for the NGC 6907 rotation curve. Some emission lines are superimposed on a broader absorption profile, which suggests that they were not formed in NGC 6908. Finally, the H I profile exhibits details of the interaction, showing three components: one for NGC 6908, another for the excited gas in the NGC 6907 disc and a last one for the gas with higher relative velocities left behind NGC 6908 by dynamical friction, used to estimate the time when the interaction started in (3.4 +/- 0.6) x 10(7) yr ago.

MODELING VERY LONG BASELINE INTERFEROMETRIC IMAGES WITH THE CROSS-ENTROPY GLOBAL OPTIMIZATION TECHNIQUE

We present a new technique for obtaining model fittings to very long baseline interferometric images of astrophysical jets. The method minimizes a performance function proportional to the sum of the squared difference between the model and observed images. The model image is constructed by summing N(s) elliptical Gaussian sources characterized by six parameters: two-dimensional peak position, peak intensity, eccentricity, amplitude, and orientation angle of the major axis. We present results for the fitting of two main benchmark jets: the first constructed from three individual Gaussian sources, the second formed by five Gaussian sources. Both jets were analyzed by our cross-entropy technique in finite and infinite signal-to-noise regimes, the background noise chosen to mimic that found in interferometric radio maps. Those images were constructed to simulate most of the conditions encountered in interferometric images of active galactic nuclei. We show that the cross-entropy technique is capable of recovering the parameters of the sources with a similar accuracy to that obtained from the very traditional Astronomical Image Processing System Package task IMFIT when the image is relatively simple (e. g., few components). For more complex interferometric maps, our method displays superior performance in recovering the parameters of the jet components. Our methodology is also able to show quantitatively the number of individual components present in an image. An additional application of the cross-entropy technique to a real image of a BL Lac object is shown and discussed. Our results indicate that our cross-entropy model-fitting technique must be used in situations involving the analysis of complex emission regions having more than three sources, even though it is substantially slower than current model-fitting tasks (at least 10,000 times slower for a single processor, depending on the number of sources to be optimized). As in the case of any model fitting performed in the image plane, caution is required in analyzing images constructed from a poorly sampled (u, v) plane.

An automatic correction method for the heel effect in digitized mammography images

The most significant radiation field nonuniformity is the well-known Heel effect. This nonuniform beam effect has a negative influence on the results of computer-aided diagnosis of mammograms, which is frequently used for early cancer detection. This paper presents a method to correct all pixels in the mammography image according to the excess or lack on radiation to which these have been submitted as a result of the this effect. The current simulation method calculates the intensities at all points of the image plane. In the simulated image, the percentage of radiation received by all the points takes the center of the field as reference. In the digitized mammography, the percentages of the optical density of all the pixels of the analyzed image are also calculated. The Heel effect causes a Gaussian distribution around the anode-cathode axis and a logarithmic distribution parallel to this axis. Those characteristic distributions are used to determine the center of the radiation field as well as the cathode-anode axis, allowing for the automatic determination of the correlation between these two sets of data. The measurements obtained with our proposed method differs on average by 2.49 mm in the direction perpendicular to the anode-cathode axis and 2.02 mm parallel to the anode-cathode axis of commercial equipment. The method eliminates around 94% of the Heel effect in the radiological image and the objects will reflect their x-ray absorption. To evaluate this method, experimental data was taken from known objects, but could also be done with clinical and digital images.