Catálogo de publicaciones - libros

Methods of tissue classification in MRI brain images play a significant role in computational neuroanatomy, particularly in automated ROI-based volumetry. A well-known and very simple k-NN classifier is used here without the need for user input during the training process. The classifier is trained with the use of tissue probability maps which are available in selected digital atlases of brain. The influence of misalignement between images and the tissue probability maps on the classifier’s efficiency is studied in this paper. Deformable registration is used here to align the images and maps. The classifier’s efficiency is tested in an experiment with data obtained from standard Simulated Brain Database.

In recent years many various biomedical image segmentation methods have appeared. Though typically presented to be successful the majority of them was not properly tested against ground truth images. The obvious way of testing the quality of new segmentation was based on visual inspection by a specialist in the given field. The novel 3D biomedical image data simulator is presented in this paper. It offers the results of high quality. The comparison of generated synthetic data is compared against real image data using standard similarity techniques.

We propose a framework for fast and automated initialization of segmentation algorithms in Computed Tomography images. Based on the idea that time-consuming voxel classification should be done only on spatially constrained areas, we build classifiers at body and slice levels which quickly define a constrained region of interest. Voxel classification is then performed by a divide-and-conquer strategy using a probabilistic-boosting tree. In addition, this framework can incorporate additional information on the volume, if available, such as the position of another organ to improve its accuracy and robustness. The framework is applied to seed extraction in kidneys and liver.

Computer-based interactive systems and robots have become a massive technology for improving human-impaired social interaction, especially for children with autistic. Autism is a lifelong development disability, often accompanied by learning technologies. As a result, they have trouble interacting within our complex social environment and are, for the most part, unable to recognize other people’s behaviors. In this paper, we present game-based therapeutic environments for people diagnosed with a mild form of autism. The proposed interactive system traces a child’s emotion with intensity for changing a game environment for the purpose of triggering their emotions. The pedagogical agent provides therapy instruction with motivational support to children through adapting a child’s emotional behaviors.

A novel approach towards a simplified, though still reliable human vision model based on the spatial masking properties of the human visual system (HVS) is presented. The model contains two basic characteristics of the HVS, namely texture masking and luminance masking. These masking effects are implemented as simple spatial filtering followed by a weighting function, and are efficiently combined into a single visibility coefficient. This HVS model is applied to a blockiness metric by using its output to scale the block-edge strength. To validate the proposed model, its performance in the blockiness metric is determined by comparing it to the same blockiness metric having different HVS-based models embedded. The results show that the proposed model is indeed simple, without compromising its accuracy.

Recent investigations into the effect of age on face identification concluded that it was more difficult to identify younger people than older ones. The identification rates of the different age groups were, however, not measured under identical conditions. There was a significantly higher percentage of younger people in all the face image samples. We found that a person from any age group will find that they look more similar to another person from the same age group, as opposed to someone from another age group. The experiments we carried out using samples that have an evenly distributed age range did not show a statistically significant difference between the sample age groups.

This paper presents a speaker-independent audio-visual digit recognition system that utilizes speech and visual lip signals. The extracted visual features are based on line-motion estimation obtained from video sequences with low resolution (128 ×128 pixels) to increase the robustness of audio recognition. The core experiments investigate lip motion biometrics as stand-alone as well as merged modality in speech recognition system. It uses Support Vector Machines, showing favourable experimental results with digit recognition featuring 83% to 100% on the XM2VTS database depending on the amount of available visual information.

Recent advances in information security requires randomly selected strong keys. Most of these keys are generated by software-based random number generators. However, implementing a Truly Random Number Generator (TRNG) without using a hardware-supported platform is not reliable. In this paper, a fingerprint authentication system using a hardware-based TRNG to produce a private key that encrypts the fingerprint template of a person is presented. The designed hardware can easily be mounted on a standard or embedded PC via its PCI interface to produce random number keys. Random numbers forming the private key is guaranteed to be true because it passes a two-level randomness test evaluated first on the FPGA then on the PC by applying the full NIST test suite. The whole system implements an AES-based encryption scheme to store the person’s secret stored on a smart or glossary card safely. The main contribution of the work is the use of new-generation hardware-based TRNGs to enhance the security of a fingerprint authentication system.

In this paper, we propose a new manifold representation capable of being applied for visual speech recognition. In this regard, the real time input video data is compressed using Principal Component Analysis (PCA) and the low-dimensional points calculated for each frame define the manifolds. Since the number of frames that from the video sequence is dependent on the word complexity, in order to use these manifolds for visual speech classification it is required to re-sample them into a fixed number of keypoints that are used as input for classification. In this paper two classification schemes, namely the k Nearest Neighbour (kNN) algorithm that is used in conjunction with the two-stage PCA and Hidden-Markov-Model (HMM) classifier are evaluated. The classification results for a group of English words indicate that the proposed approach is able to produce accurate classification results.

Since fingerprints provide a reliable alternative for traditional password based security systems, they gain industry and citizen acceptance. However, due to the higher uncertainty and inherent complexity associated with biometrics, using pure biometric traits does not present a reliable security system especially for large populations. This paper addresses this problem by proposing a hardening scheme which combines the fingerprint minutiae-based template and user-specific pseudo random data to enhance security. In the proposed scheme, a set of randomly selected user-specific chaff minutiae features are stored in a smartcard and a subset of this set is used at each acquisition. The set of chaff minutiae is combined with the template set and scrambled to form a fixed-length hardened feature. The graph based dynamic matching algorithm is transparent to the proposed hardening scheme anyhow it runs as if pure original template and query features are used. Our experiments show that biometric hardening reduces error rate to 0% with several orders of magnitude separation between genuine and impostor populations.