The integral image representation of an image is important for a large number of modern image processing algorithms. Integral image representations can reduce computation and increase the operating speed of certain algorithms, improving real-time performance. Due to increasing demand for real-time image processing performance, an integral image architecture capable of accelerating the calculation based on the amount of available resources is presented. Use of the proposed accelerator allows for subsequent stages of a design to have data sooner and execute in parallel. It is shown here how, with some additional resources used in the Field Programmable Gate Array (FPGA), a speed increase is obtained by using a one-dimensional Systolic Array (SA) approach. Additionally, extra guidelines are given for further research in this area.

This work deals with task of face detection on graphic card. First part is the introduction to face detection methods focusing on detector proposed by Viola and Jones. Further, this work studies the possibilities of mapping detector's key parts on graphic card. Next part describes implementation details of designed application. The end of work include results and comparison with CPU approach. The last chapter summarizes the whole work and proposes future possibilities of development.

There is a rising interest in geospatial object detection due to not only the complexity of manual processing of such huge amount of data provided by high resolution satellite imagery but also for military application needs. A fundamental and yet state-of-the art approach for object detection is based on methods that utilize the global appearance. In such a holistic approach, the information of the object class is aimed to be modeled as a whole in the learning phase. And during the classification, a decision is taken at each window of the test image. In this thesis, two different discriminative methods are investigated for airplane detection from satellite images. In the first method, Haar-like features are used as weak classifiers for the airplane class representation. Then the AdaBoost learning algorithm is used to select the critical visual features that represent the airplanes best. Finally, a cascade of classifiers is constructed in order to speed-up the classifier. In the second method, a computationally efficient appearance-based algorithm for airplane detection is presented. An operator exploiting the edge information via gray level differences between the target and its background is constructed with Haar-like polygon regions using the shape information of the airplane as an invariant. The airplanes matching the operator are supposed to yield higher responses around the centroid of the object. Fast evaluation of the operator is achieved by means of integral image. The proposed algorithm has promising results in terms of accuracy in detecting aircraft type geospatial objects from satellite imagery.

This MSc Thesis deals with face detection in image. In this approach, facial features (eyes, nose, mouth corners) are detected first and then joined to the whole face. For the facial features detection, classifiers trained with AdaBoost algorithm are used. Haar wavelets are used as features for classification.

No description available.

Production of artificial Three-Dimension (3-D) images was the aim of many researches over hundreds of years. 3-D images are the images that create sense of depth when viewing them. 3-D images are closer to the real world scenes than 2-D images due to the 3-D effect or the sense of depth the 3-D images provide. Sense of depth can be caused by binocular cues including convergence and parallax. Convergence is created by the difference between the angles of the left eye and the right eye viewing axes. Parallax is the effect of viewing with one eye a view of the scene that is inherently shifted to the view seen by the other eye. Several techniques have targeted the creation of 3-D images with the mentioned cues. The technique is preferred when it is able to create 3-D images so that the viewer can view these images without wearing special glasses and the occurrence of viewer fatigue. Integral photography that was invented in 1908 is able to meet the previous requirements. Based on integral photography, several techniques, research and studies have been published. The purposes of this thesis include the computer simulation of flexible integral photography systems, the computer generation of good quality 3-D static and animated integral images using the simulated systems, optimising the generation process to be more accurate, less expensive, more effective, and faster, and producing a portable specialist software tool to achieve these targets. New techniques and algorithms are needed to meet these purposes. A literature survey was carried out about the closest researches and studies to the subject of computer-generated integral images; these were compared with the new techniques introduced in this study to prove the advantages and the necessity of these new techniques. The closest technique to the suggested techniques was implemented using more developed tools to compare the quality of the resulting integral images with the targeted integral images that are going to be produced using the tools and algorithms proposed in this thesis. A method to simulate an imaging system and produce integral images based on the new technique of dividing the view volume of the scene was introduced, explained, proved, and implemented with a program designed for this purpose. To optimise the processing time and the image quality, the previous method is developed, new features are added to the resulting integral images, and better performance was achieved by introducing the method of Displacing the Virtual Camera Target (DCT). Application software with Graphical User Interface is designed and implemented to allow users to select the required parameters of the imaging system and the required features of the resultant integral images. The software tool that is based on the developed techniques and employing OpenGL is useful to simulate the imaging systems, tune their parameters before the actual implementation of these systems, and as a result, save time and materials when designing these systems. The introduced techniques and the software tools are faster, more effective, and cheaper original methods to help in optimising both the integral imaging systems and the quality of integral images. These software tools based on the new techniques can be used on a wide range of devices and platforms because these are employing the portable Application Interface OpenGL. With these methods, integral imaging systems are simulated, and optimised; good quality static and animated integral images were created.

This thesis deals with a face detection on mobile phones. It especially focuses on Windows Mobile platform. The introduction is therefore devoted to this operating system and alternatives of working with the camera. The next part of the text refers to general problems of the face detection in the image considering the weak performance of the target device. Another part of this thesis is a description of the acquisition of images from the camera using DirectShow multimedia framework and creation of a custom transformation filter for the face detection. Achieved results are summarized in the conclusion. It takes a form of tests examining different mobile devices. All difficulties arising during Windows Mobile developing are also mentioned.

This work focuses on classification and recognition of traffic signs in image. It describes briefly some used methods a deeply describes chosen system including extensions and method for creating models needed for classification. There's described implementation of library and demonstration program including important pieces of knowledge discovered during development. There're also results of some experiments and possible enhancements in conclusion.

This paper describes methods to track a user-defined point in the vision of a robot as it drives forward. This tracking allows a robot to keep itself directed at that point while driving so that it can get to that user-defined point. I develop and present two new multi-scale algorithms for tracking arbitrary points between two frames of video, as well as through a video sequence. The multi-scale algorithms do not use the traditional pyramid image, but instead use a data structure called an integral image (also known as a summed area table). The first algorithm uses edge-detection to track the movement of the tracking point between frames of video. The second algorithm uses a modified version of the Moravec operator to track the movement of the tracking point between frames of video. Both of these algorithms can track the user-specified point very quickly. Implemented on a conventional desktop, tracking can proceed at a rate of at least 20 frames per second.

Master´s thesis deals with current problems of face detection and its parts in the infrared thermographic spectrum. Most previously published literature deals with the detection in the visible spectrum, making the thermographic detection range an interesting alternative. The work deals with the processing of image signals, images and faces in thermographic spectrum, selected methods of face detection and its parts and also deals with practical system design for detecting facial parts in this spectrum and its subsequent testing.