<p>Face detection has been a research area for more than ten years. It is a complex problem due to the high variability in faces and amongst faces; therefore it is not possible to extract a general pattern to be used for detection. This is what makes the face detection problem a challenge.</p><p>This thesis gives the reader a background to the face detection problem, where the two main approaches of the problem are described. A face detection algorithm is implemented using a context-based method in combination with an evolving neural network. The algorithm consists of two majors steps: detect possible face areas and within these areas detect faces. This method makes it possible to reduce the search space.</p><p>The performance of the algorithm is evaluated and analysed. There are several parameters that affect the performance; the feature extraction method, the classifier and the images used.</p><p>This work resulted in a face detection algorithm and the performance of the algorithm is evaluated and analysed. The analysis of the problems that occurred has provided a deeper understanding for the complexity of the face detection problem.</p>

The ability to detect faces in images is of critical ecological significance. It is a pre-requisite for other important face perception tasks such as person identification, gender classification and affect analysis. Here we address the question of how the visual system classifies images into face and non-face patterns. We focus on face detection in impoverished images, which allow us to explore information thresholds required for different levels of performance. Our experimental results provide lower bounds on image resolution needed for reliable discrimination between face and non-face patterns and help characterize the nature of facial representations used by the visual system under degraded viewing conditions. Specifically, they enable an evaluation of the contribution of luminance contrast, image orientation and local context on face-detection performance.

Face detection has been a research area for more than ten years. It is a complex problem due to the high variability in faces and amongst faces; therefore it is not possible to extract a general pattern to be used for detection. This is what makes the face detection problem a challenge. This thesis gives the reader a background to the face detection problem, where the two main approaches of the problem are described. A face detection algorithm is implemented using a context-based method in combination with an evolving neural network. The algorithm consists of two majors steps: detect possible face areas and within these areas detect faces. This method makes it possible to reduce the search space. The performance of the algorithm is evaluated and analysed. There are several parameters that affect the performance; the feature extraction method, the classifier and the images used. This work resulted in a face detection algorithm and the performance of the algorithm is evaluated and analysed. The analysis of the problems that occurred has provided a deeper understanding for the complexity of the face detection problem.

Face detection is a classical application of object detection. There are many practical applications in which face detection is the first step; face recognition, video surveillance, image database management, video coding. This report presents the results of an implementation of the AdaBoost algorithm to train a Strong Classifier to be used for face detection. The AdaBoost algorithm is fast and shows a low false detection rate, two characteristics which are important for face detection algorithms. The application is an implementation of the AdaBoost algorithm with several command-line executables that support testing of the algorithm. The training and detection algorithms are separated from the rest of the application by a well defined interface to allow reuse as a software library. The source code is documented using the JavaDoc-standard, and CppDoc is then used to produce detailed information on classes and relationships in html format. The implemented algorithm is found to produce relatively high detection rate and low false alarm rate, considering the badly suited training data used.

The digital image processing has been developed for a long time. The image detection and tracking are involved to a variety of digital techniques. In this research we introduce the digital image processing techniques, base on a boosted cascade of simple features to develop a face detection and tracking system. Due to a large amount of computation in face detection under the complex environment will affect the detection rate and velocity efficiency. Therefore, we use the extended feature and set of 45ﾟ rotated feature with fast feature computation which called the integral image, combine with the deformable templates. We can compute a part of the image block to reduce the computation and improve the system. In the PAN-TILT unit, we use fuzzy logic. The results of experiment show that system is robust and fast.

Face detection has many applications in technical area. We think about accuracy and regular arrangement of data of face detection. So, we select Recognition algorithms using neural network for implementation. The implementation method can be divided into three parts. One is Modified Census Transform. The other one is computing hypotheses. Other is square frame for mark face. Modified Census Transform is a regularly computing method and regular arrangement of data. Modified Census Transform is compatible using SIMD execution, but other parts is irregular arrangement of data and not easy to parallel execution. This paper uses SIMD processor architecture which develops in our laboratory to implementation of Modified Census Transform and multi-data streaming property. The picture is divided four parts to execute at the same time and changes different mode to execute according to different algorithm then fetch data is smooth and moving data can reduce frequency. Adding a new instruction that uses 16bits data format uses four MMX registers for 4¡Ñ4 transpose of the matrix. The other is loading data and extending signed bit or unsigned bit at the same time. They can accelerate parallel execution in multi-data streaming. We also support multi-data streaming that is not series. It uses striping mode to fetch multi-data which between the same distance then we can achieve to compute multi-data streaming. Besides, we use hypotheses to distinguish different person that we only want find one. We compare two hypotheses. If the difference in hypotheses between two different picture that there is small than 0.3%, they are the same person which in different picture. Finial, we verify the function is correct in UMVP-2500 platform. We compare efficiency with MMX and Xscale and analysis multi-data streaming SIMD architecture which has some benefits. We compare efficiency with MMX. We speed up 373%. We compare efficiency with Xscale. We speed up 345%. This result will show that multi-data streaming SIMD architecture compares speed up with others SIMD architecture. Multi-data streaming SIMD architecture adds a new instruction which is 4¡Ñ4 transpose of the matrix. Because the 4¡Ñ4 transpose of the matrix can change row and column, we have new abstraction. The common computation likes a line, but the new abstraction becomes a phase. MMX and Xscale are not this abstraction.

Human face detection and recognition techniques play an important role in applica- tions like face recognition, video surveillance, human computer interface and face image databases. Using color information in images is one of the various possible techniques used for face detection. The novel technique used in this project was the combination of various techniques such as skin color detection, template matching, gradient face de- tection to achieve high accuracy of face detection in frontal faces. The objective in this work was to determine the best rotation angle to achieve optimal detection. Also eye and mouse template matching have been put to test for feature detection.

In this paper, we introduce a novel set of features for robust object recognition, which exhibits outstanding performances on a variety ofobject categories while being capable of learning from only a fewtraining examples. Each element of this set is a complex featureobtained by combining position- and scale-tolerant edge-detectors overneighboring positions and multiple orientations.Our system - motivated by a quantitative model of visual cortex -outperforms state-of-the-art systems on a variety of object imagedatasets from different groups. We also show that our system is ableto learn from very few examples with no prior category knowledge. Thesuccess of the approach is also a suggestive plausibility proof for aclass of feed-forward models of object recognition in cortex. Finally,we conjecture the existence of a universal overcompletedictionary of features that could handle the recognition of all objectcategories.

Abstract Automatic face recognition is a challenging problem which has received much attention during recent years due to its many applications in different fields such as law enforcement, security applications, human-machine interaction etc. Up to date there is no technique that provides a robust solution for all situations and different applications. From still gray images to face sequences (and passing through color images), this thesis provides new algorithms to learn, detect and recognize faces. It also analyzes some emerging directions such as the integration of facial dynamics in the recognition process. To recognize faces, the thesis proposes a new approach based on Local Binary Patterns (LBP) which consists of dividing the facial image into small regions from which LBP features are extracted and concatenated into a single feature histogram efficiently representing the face image. Then, face recognition is performed using a nearest neighbor classifier in the computed feature space with Chi-square as a dissimilarity metric. The extensive experiments clearly show the superiority of the proposed method over the state-of the-art algorithms on FERET tests. To detect faces, another LBP-based representation which is suitable for low-resolution images, is derived. Using the new representation, a second-degree polynomial kernel SVM classifier is trained to detect frontal faces in complex gray scale images. Experimental results using several complex images show that the proposed approach performs favorably compared to the state-of-art methods. Additionally, experiments with detecting and recognizing low-resolution faces are carried out to demonstrate that the same facial representation can be efficiently used for both the detection and recognition of faces in low-resolution images. To detect faces when the color cue is available, the thesis proposes an approach based on a robust model of skin color, called a skin locus, which is used to extract the skin-like regions. After orientation normalization and based on verifying a set of criteria (face symmetry, presence of some facial features, variance of pixel intensities and connected component arrangement), only facial regions are selected. To learn and visualize faces in video sequences, the recently proposed algorithms for unsupervised learning and dimensionality reduction (LLE and ISOMAP), as well as well known ones (PCA, SOM etc.) are considered and investigated. Some extensions are proposed and a new approach for selecting face models from video sequences is developed. The approach is based on representing the face manifold in a low-dimensional space using the Locally Linear Embedding (LLE) algorithm and then performing K-means clustering. To analyze the emerging direction in face recognition which consists of combining facial shape and dynamic personal characteristics for enhancing face recognition performance, the thesis considers two factors (face sequence length and image quality) and studies their effects on the performance of video-based systems which attempt to use a spatio-temporal representation instead of a still image based one. The extensive experimental results show that motion information enhances automatic recognition but not in a systematic way as in the human visual system. Finally, some key findings of the thesis are considered and used for building a system for access control based on detecting and recognizing faces.

In the field of computer vision, face detection concerns the positive identification of the faces of people within still images or video streams, which is extremely useful for applications such as counting, tracking, and recognition. When applied in large-scale environments, such as lecture theatres, we have found that existing technology can struggle greatly in detecting faces due primarily to the indiscernibility of their features, caused by partial occlusion, problematic orientation, and a lack of focus or resolution. We attempt to overcome this issue by proposing an adaptive framework, capable of collating the results of numerous existing detection systems in order to significantly improve recall rates. This approach uses supplementary modalities, invariant to the issues posed to features, to eliminate false detections from collated sets and allow us to produce results with extremely high confidence. The properties we have selected as the bases of detection classification are size and colour, as we believe that filters that consider them can be constructed adaptively, on a per-image basis, ensuring that the variabilities inherent to large-scale imagery can be fully accounted for, and that false detections and actual faces can be accurately distinguished between on a consistent basis. The application of principal component analysis to precise face detection results yields planar size distribution models that we can use to discard results that are either too large or too small to realistically represent faces within given images. Classifying a detection according to the correspondence of its general colour tone to the expected colour of skin is a more complex matter, however, as the apparent colour of skin is highly dependent upon incident illumination, and existing techniques are neither specific nor flexible enough to model it as accurately as we believe possible. Therefore, we propose another system, which will be able to adaptively model skin colour distributions according to the Gaussian probability densities exhibited by the colours of precise face detections. Furthermore, it will be suitable for independent application to real-time skin segmentation tasks as a result of considerable optimisation. This thesis details the design, the development, and the implementation of our systems, and thoroughly evaluates them with regards to the accuracy of their results and the efficiency of their performances, thereby establishing fully the suitability of them for solving certain types of presented problems.