Electronic Letters on Computer Vision and Image Analysis (ELCVIA - Universitat Autònoma de Barcelona)
Not a member yet
343 research outputs found
Sort by
Random Image Matching CAPTCHA System
Full text linkSecurity risks is an important issues and caught the attention of researchers in the area of networks, web development, human computer interaction and software engineering. One main challenge for online systems is to identify whether the users are humans or software robots (bots). While it is natural to provide service to human users, providing service for software robots (bots) comes with many security risks and challenges. Software robots are often used by spammers to create fake online accounts, affect search engine ranking, take part in on-line polls, send out spam or simply waste the resources of the server. In this paper we introduce a visual CAPTCHA technique that is based on generating random images by the computer, theuser is then asked to match a feature point between two images (i.e. solve the correspondence problem as defined by the researchers in the computer vision area). The relationship between the two images is based on a randomly generated homography transformation function. The main advantage of our approach compared to other visual CAPTCHA techniques is that we eliminate the need for a database of images while retaining ease of use
Design and Stability Analysis of Multi-Objective Ensemble Classifiers
Full text linkSome important topics, which affects directly on the performance of the designed ensemble classifier, inflict a complex search space with high dimensions on the researcher. So, heuristic algorithms can be applied to find best solutions because of their capability of efficient search in the solution space. Due to the stochastic nature of heuristic algorithms, it\u27s necessary to perform stability analysis of heuristic ensemble classifiers. In this paper, Multi-Objective Inclined Planes Optimization (MOIPO) algorithm, as a novel multi-objective technique, is used to design ensemble classifiers and the performance of created ensemble is compared with ensemble designed by Multi-Objective Particle Swarm Optimization (MOPSO) algorithm.Experimental results confirm the supremacy of MOIPO for designing ensemble classifiers. So, in the next step, for the first time, the stability of this ensemble classifier is analyzed by using statistical method and suitable model for stability analysis is specified
Detection of retinal blood vessels from ophthalmoscope images using morphological approach
Full text linkAccurate segmentation of retinal blood vessels is an essential task for diagnosis of various pathological disorders. In this paper, a novel method has been introduced for segmenting retinal blood vessels which involves pre-processing, segmentation and post-processing. The pre-processing stage enhanced the image using contrast limited adaptive histogram equalization and 2D Gabor wavelet. The enhanced image is segmented using geodesic operators and a final segmentation output is obtained by applying a post-processing stage that involves hole filling and removal of isolated pixels. The performance of the proposed method is evaluated on the publicly available Digital retinal images for vessel extraction (DRIVE) and High-resolution fundus (HRF) databases using five different measurements and experimental analysis shows that the proposed method reach an average accuracy of 0.9541 on DRIVE database and 0.9568, 0.9478 and 0.9613 on HRF database with healthy, diabetic retinopathy (DR) and glaucomatous images respectively
An ant colony based model to optimize parameters in industrial vision
Full text linkIndustrial vision constitutes an efficient way to resolve quality control problems. It proposes a wide variety of relevant operators to accomplish controlling tasks in vision systems. However, the installation of these systems awaits for a precise parameter tuning, which remains a very difficult exercise. The manual parameter adjustment can take a lot of time, if precision is expected, by revising many operators. In order to save time and get more precision, a solution is to automate this task by using optimization approaches (mathematical models, population models, learning models...). This paper proposes an Ant Colony Optimization (ACO) based model. The process considers each ant as a potential solution, and then by an interacting mechanism, ants converge to the optimal solution. The proposed model is illustrated by some image processing applications giving very promising results. Compared to other approaches, the proposed one is very hopeful.
Indoor/outdoor navigation system based on possibilistic traversable area segmentation for visually impaired people
Full text linkAutonomous collision avoidance for visually impaired people requires a specific processing for an accurate definition of traversable area. Processing of a real time image sequence for traversable area segmentation is quite mandatory. Low cost systems suggest use of poor quality cameras. However, real time low cost camera suffers from great variability of traversable area appearance at indoor as well as outdoor environments. Taking into account ambiguity affecting object and traversable area appearance induced by reflections, illumination variations, occlusions (, etc...), an accurate segmentation of traversable area in such conditions remains a challenge. Moreover, indoor and outdoor environments add additional variability to traversable areas. In this paper, we present a real-time approach for fast traversable area segmentation from image sequence recorded by a low-cost monocular camera for navigation system. Taking into account all kinds of variability in the image, we apply possibility theory for modeling information ambiguity. An efficient way of updating the traversable area model in each environment condition is to consider traversable area samples from the same processed image for building its possibility maps. Then fusing these maps allows making a fair model definition of the traversable area. Performance of the proposed system was evaluated on public databases, with indoor and outdoor environments. Experimental results show that this method is challenging leading to higher segmentation rates
Edge-aware wedgelet estimation for depth maps compression
Full text linkIn recent years, Multi-view Video plus Depth (MVD) compression has received much attention thanks to its relevance to free viewpoint applications needs. An efficient compression, that causes the least distortion without excessive rate and complexity increase, becomes a must particularly for depth maps. These latter can be compressed efficiently by the 3D extension of High Efficiency Video Coding (3D-HEVC), which has explored wedgelets. Such functions lead to significant Rate-Distortion tradeoffs. However, they require a very large computational complexity involved by the exhaustive search used for the estimation of the wedgelet subdivision line. In this paper, we propose a rapid localization of this latter using an edge detection approach. The experimental results show that the proposed approach allows an important gain in terms of encoding delay, while providing better depth maps and synthesized views quality compared to the exhaustive search approach
Efficient Poisson Image Editing
Full text linkImage composition refers to the process of composing two or more images to create an acceptable output image. It is one of the important techniques of image processing. In this paper, two efficient methods for composing color images are proposed. In the proposed methods, the Poisson equation is solved using image pyramid, and divide-and-conquer methods. The proposed methods are more efficient than other existing image composition methods. They reduce the time taken in the composition process while achieving almost identical results using the previous image composition methods. In the proposed methods, the Poisson equation is solved after converting it to a linear system using different methods. The results show that the time for composing color images is decreased using the proposed methods
Monitoring Infants by Automatic Video Processing
Full text linkThis work has, as its objective, the development of non-invasive and low-cost systems for monitoring and automatic diagnosing specific neonatal diseases by means of the analysis of suitable video signals. We focus on monitoring infants potentially at risk of diseases characterized by the presence or absence of rhythmic movements of one or more body parts. Seizures and respiratory diseases are specifically considered, but the approach is general.Seizures are defined as sudden neurological and behavioural alterations. They are age-dependent phenomena and the most common sign of central nervous system dysfunction. Studies indicate an incidence rate of neonatal seizures of 2‰ for live births, 11‰ for preterm neonates, and 13‰ for infants weighing less than 2500 g at birth. Seizures in newborns have to be promptly and accurately recognized in order to establish timely treatments that could avoid an increase of the underlying brain damage.Respiratory diseases related to the occurrence of apnoea episodes may be caused by cerebrovascular events. Among the wide range of causes of apnoea, besides seizures, a relevant one is Congenital Central Hypoventilation Syndrome (CCHS). With a reported prevalence of 1 in 200,000 live births, CCHS, formerly known as Ondine’s curse, is a rare life-threatening disorder characterized by a failure of the automatic control of breathing, caused by mutations in a gene classified as PHOX2B. The reported mortality rates range from 8% to 38% of newborn with genetically confirmed CCHS. Nowadays, CCHS is considered a disorder of autonomic regulation, with related risk of sudden infant death syndrome (SIDS).Currently, the standard method of diagnosis, for both diseases, is based on polysomnography, a set of sensors such as ElectroEncephaloGram (EEG) sensors, ElectroMyoGraphy (EMG) sensors, ElectroCardio-Graphy (ECG) sensors, elastic belt sensors, pulse-oximeter and nasal flow-meters. This monitoring system is very expensive, time-consuming, moderately invasive and requires particularly skilled medical personnel, not always available in a Neonatal Intensive Care Unit (NICU). Therefore, automatic, real-time and noninvasive monitoring equipments able to reliably recognize these diseases would be of significant value in the NICU.A very appealing monitoring tool to automatically detect neonatal seizures or breathing disorders may be based on acquiring, through a network of sensors, e.g., a set of video cameras, the movements of the newborn’s body (e.g., limbs, chest) and properly processing the relevant signals. An automatic multi-sensor system could be used to permanently monitor every patient in the NICU or specific patients at home. Furthermore, a wire-free technique may be more user-friendly and highly desirable when used with infants, in particular with newborns. We have focused on a reliable method to estimate the periodicity in pathological movements based on theuse of the Maximum Likelihood (ML) criterion. In particular, average differential luminance signals from multiple sensors are extracted and the presence or absence of a significant periodic component is analysed in order to detect possible pathological conditions. Analysis of the data obtained from multiple sensors placed around a patient, makes it possible to increase the reliability of the detection system. This approach is very versatile and allowed us to investigate various scenarios, including: a single RGB camera, an RGB-Depth sensor and a network of a few RGB cameras. Data fusion principles are considered to aggregate the signals from multiple sensors. In the case of respiratory diseases, since chest movements are subtle, the video can be pre-processed by a recently proposed selective magnification algorithm, namely the eulerian video magnification (EVM), which has the purpose of emphasizing small movements. Within this context, we have also developed a second improved algorithm in order to speed up the processing time required for the detection of apnoeas, limiting the computational load. Moreover, in order to have, at any time, a subject on which to test the continuously evolving detection algorithms, we have decided to realize two low-cost programmable simulators able to replicate the symptomatic movements characteristic of the diseases under consideration.The performance of the proposed detection algorithms is assessed, in terms of sensitivity, specificity, and Receiver Operating Characteristic (ROC) curves, considering real video recordings of newborns provided by a Neonatal Intensive Care Unit (NICU). The diagnostic performance of our detection systems has been compared to that of the gold standard based on a prolonged polysomnographic EEG monitoring. It is important to stress how we have always pursued simplicity, because low complexity leads to a low processing time, and this means that these algorithms can be used on a wide range of hardware devices. In particular, we have developed a smartphone App, called “Smartphone based contactless epilepsy detector” (SmartCED), able to detect neonatal clonic seizures and warn the user about their occurrence in real-time. With this powerful inexpensive monitoring system every child, or adult, could be easily monitored at home without additional hardware costs. SmartCED is designed for an easy and intuitive utilization, although it integrates complex software. The App presents, indeed, a user-friendly interface in order to extend its use to even unskilled staff. The user has to start the App, frame the patient and start monitoring the patient with a simple touch
Statistical and deterministic approaches for multimedia forensics
Full text linkThe increasing availability and pervasiveness of multimedia data in our society is before our very eyes.As a result of globalization and worldwide connectivity, people from all over the planet are exchanging constantly increasing amounts of images, videos, audio recordings on a daily basis. Coupled with the easy access to user-friendly editing software, this poses a number of problems related to the reliability and trustworthiness of such content, as well as its potential malevolent use. For this reason, the research field of multimedia forensics focuses on the development of forensic tools for verifying the authenticity of multimedia data. The hypothesis of pristine status of images, videos or audio tracks is called into question and can be rejected if traces of manipulation are detected with a certain degree of confidence. In this framework, studying traces left by any operation that could have been employed to process the data, either for malicious purposes or simply to improve their content or presentation, turns out to be of interest for a comprehensive forensic analysis. The goal of this doctoral study is to contribute to the field of multimedia forensics by exploiting intrinsic statistical and deterministic properties of multimedia data. With this respect, much work has been devoted to the study of JPEG compression traces in digital images, resulting in the development of several innovative approaches. Indeed, some of the main related research problems have been addressed and solution based on statistical properties of digital images have been proposed. In particular, the problem of identifying traces of JPEG compressions in images that have been decompressed and saved in uncompressed formats has been extensively studied, resulting in the design of novel statistical detectors. Given the enormous practical relevance, digital images in JPEG formats have also been considered. A novel method aimed at discriminating images compressed only once and more than once has been developed, and tested on a variety of images and forensic scenarios. Being the potential presence of intelligent counterfeiters ever increasingly studied, innovative counterforensic techniques to JPEG compression based on smart reconstruction strategies are proposed.Finally, we explore the possibility of defining and exploiting deterministic properties related to a certain processing operation in the forensic analysis. With this respect, we present a first approach targeted to the detection in one-dimensional data of a common data smoothing operation, the median filter. A peculiarity of this method is the ability of providing a deterministic response on the presence of median filtering traces in the data under investigation
Hierarchical Visual Content Modelling and Query based on Trees
Full text linkIn recent years, such vast archives of video information have become available that human annotation of content is no longer feasible; automation of video content analysis is therefore highly desirable. The recognition of semantic content in images is a problem that relies on prior knowledge and learnt information and that, to date, has only been partially solved. Salient analysis, on the other hand, is statistically based and highlights regions that are distinct from their surroundings, while also being scalable and repeatable. The arrangement of salient information into hierarchical tree structures in the spatial and temporal domains forms an important step to bridge the semantic salient gap.Salient regions are identified using region analysis, rank ordered and documented in a tree for further analysis. A structure of this kind contains all the information in the original video and forms an intermediary between video processing and video understanding, transforming video analysis to a syntactic database analysis problem.This contribution demonstrates the formulation of spatio-temporal salient trees the syntax to index them, and provides an interface for higher level cognition in machine vision