A Robust Color Circle-Marker Detection Algorithm based on Color information and Hough Transformation

International audience

Ridge-line optimal detector

Image processing techniques have seen many developments in recent years. Starting from the pioneering work of Canny, Deriche developed a second order recursive filter capable of detecting stepped contours. However, there are other contour shapes that those filters struggle to detect. We describe a new optimal filter sensu Canny for detecting ridge-line contours. This is a third order recursive and even filter. It is dependent on three parameters by which detection accuracy is adjusted. The results obtained by applying this filter to (possibly noise- affected) images are compared with those in the work by Ziou. © 2000 Society of Photo-Optical Instrumentation Engineers. (S0091-3286(00)00706-6)

Image enhancement in simple fingerprint minutiae extraction algorithm using crossing number on valley structure

In fingerprint recognition system, fingerprint feature extraction algorithm requires good quality fingerprint images to produce good results. Therefore, one step in the preprocessing stage is image enhancement to improve the quality of poor fingerprint image, so the minutiae points can be detected with good results. In this paper, we present how this enhancement process in simple minutiae detection algorithm using crossing number on valley structure improves detection of true minutiae.

A wavelet-based demosaicking algorithm for embedded applications

This paper presents an alternative to the spatial reconstruction of the sampled color filter array acquired through a digital image sensor. A demosaicking operation has to be applied to the raw image to recover the full-resolution color image. We present a low-complexity demosaicking algorithm processing in the wavelet domain. Produced images are available at the output of the algorithm either in the spatial representation or directly in the wavelet domain for high-level post processing in the latter domain. Results show that the computational complexity has been lowered by a factor of five compared to state of the art demosaicking algorithms.

Generalization of Canny–Deriche filter for detection of noisy exponential edge

This paper presents a generalization of the Canny-Deriche filter for ramp edge detection with optimization criteria used by Canny (signal-to-noise ratio, localization, and suppression of false responses). Using techniques similar to those developed by Deriche, we derive a filter which maximizes the product of the first two criteria under the constraint of the last one. The result is an infinite length impulse response filter which leads to a stable third-order recursive implementation. Its performance shows an increase of the signal-to-noise ratio in the case of blurred and noisy images, compared to the results obtained from Deriche's filter.

Geometric deformation measurement and correction applied to dynamic streak camera images

The complete procedure of measuring and correcting geometric deformations encountered with dynamic streak camera images in the picosecond range is presented and discussed. First, we describe the experimental setup derived from the well known spacing calibration grid method. The implemented measurement bench, adapted to time-resolved 1D imaging, notably exhibits a great accuracy and repeatability both in space and time thanks to a three-axis motorized translation stage and programmable delay lines. Second, we examine image restoration by two different analytical transform means (local versus global): results and performances of both are compared. Then we deal with final image reconstruction …

Real-Time 3D Face Acquisition Using Reconfigurable Hybrid Architecture

Acquiring 3D data of human face is a general problem which can be applied in face recognition, virtual reality, and many other applications. It can be solved using stereovision. This technique consists in acquiring data in three dimensions from two cameras. The aim is to implement an algorithmic chain which makes it possible to obtain a three-dimensional space from two two-dimensional spaces: two images coming from the two cameras. Several implementations have already been considered. We propose a new simple real-time implementation based on a hybrid architecture (FPGA-DSP), allowing to consider an embedded and reconfigurable processing. Then we show our method which provides depth map of …

Performance evaluation of simple fingerprint minutiae extraction algorithm using crossing number on valley structure

In fingerprint recognition system, performance of fingerprint feature extraction algorithm is important. We use visual analysis to evaluate this performance. 100 respondents fill a questionnaire consisting of 30 images from fingerprint feature extraction process. We get 12,3 % minutiae points missed by this algorithm. With BOZORTH3 minutiae matching algorithm, the distribution of matching score of 80-fingerprint images are presented and we obtain EER 5.89 % at threshold value 180.

Development of a fast panoramic face mosaicking and recognition system

We present some development results on a system that per- forms mosaicking of panoramic faces. Our objective is to study the fea- sibility of panoramic face construction in real time. To do so, we built a simple acquisition system composed of five standard cameras, which together can take simultaneously five views of a face at different angles. Then, we chose an easily hardware-achievable algorithm, consisting of successive linear transformations, in order to compose a panoramic face from these five views. The method has been tested on a large number of faces. In order to validate our system, we also conducted a preliminary study on panoramic face recognition, based on the principal-compone…

Use of wavelet for image processing in smart cameras with low hardware resources

International audience; Images from embedded sensors need digital processing to recover high-quality images and to extract features of a scene. Depending on the properties of the sensor and on the application, the designer fits together different algorithms to process images. In the context of embedded devices, the hardware supporting those applications is very constrained in terms of power consumption and silicon area. Thus, the algorithms have to be compliant with the embedded specifications i.e. reduced computational complexity and low memory requirements. We investigate the opportunity to use the wavelet representation to perform good quality image processing algorithms at a lower compu…

A 3-D stereovision simulator for centrifugal fertilizer granule spreading

International audience

Smart camera design for intensive embedded computing

Computer-assisted vision plays an important role in our society, in various fields such as personal and goods safety, industrial production, telecommunications, robotics, etc. However, technical developments are still rare and slowed down by various factors linked to sensor cost, lack of system flexibility, difficulty of rapidly developing complex and robust applications, and lack of interaction among these systems themselves, or with their environment. This paper describes our proposal for a smart camera with real-time video processing capabilities. A CMOS sensor, processor and, reconfigurable unit associated in the same chip will allow scalability, flexibility, and high performance.

Real-time flaw detection on a complex object: comparison of results using classification with a support vector machine, boosting, and hyperrectangle-based method

We present a classification work performed on industrial parts using artificial vision, a support vector machine (SVM), boost- ing, and a combination of classifiers. The object to be controlled is a coated heater used in television sets. Our project consists of detect- ing anomalies under manufacturer production, as well as in classi- fying the anomalies among 20 listed categories. Manufacturer speci- fications require a minimum of ten inspections per second without a decrease in the quality of the produced parts. This problem is ad- dressed by using a classification system relying on real-time ma- chine vision. To fulfill both real-time and quality constraints, three classification algorit…

<title>Architecture for real-time wood inspection</title>

This study has been realized to improve industrial machines that allow to analyze planks by detecting their width and too important defects thanks to a computer vision system. These machines are currently piloted by software with the help of PCs. The aim of our work is to realize a hardware card to increase the processing speed.

Computational evidence that frequency trajectory theory does not oppose but emerges from age-of-acquisition theory.

International audience; According to the age-of-acquisition hypothesis, words acquired early in life are processed faster and more accurately than words acquired later. Connectionist models have begun to explore the influence of the age/order of acquisition of items (and also their frequency of encounter). This study attempts to reconcile two different methodological and theoretical approaches (proposed by Lambon Ralph & Ehsan, 2006 and Zevin & Seidenberg, 2002) to age-limited learning effects. The current simulations extend the findings reported by Zevin and Seidenberg (2002) that have shown that frequency trajectories (FTs) have limited and specific effects on word-reading tasks. Using th…

A principal component analysis based method for the Simulation of turbulence-degraded infrared image sequence

Une methode originale de simulation d’images degradees par la turbulence atmospherique est presentee. Les methodes existantes ne permettent de simuler que des images temporellement decorrelees les unes des autres, dans le cas de l ’isoplanetisme ou du faible anisoplanetisme. Dans cet article, une simulation pour le cas de fort anisoplanetisme est proposee et une etude sur l’aspect temporel du phenomene dans le but de construire une sequence d’images degradees, a l’aide de l’analyse en composantes principales est faite. Les images obtenues montrent clairement les effets de lanisoplanetisme et l’evolution temporelle de la turbulence.

Real Time Image Rotation Using Dynamic Reconfiguration

Abstract Field programmable gate array (FPGA) components are widely used nowdays to implement various algorithms, such as digital filtering, in real time. The emergence of dynamically reconfigurable FPGAs made it possible to reduce the number of necessary resources to carry out an image-processing task (tasks chain). In this article, an image-processing application, image rotation, that exploits the FPGAs dynamic reconfiguration method is presented. This paper shows that the choice of an implementation, static or dynamic reconfiguration, depends on the nature of the application. A comparison is carried out between the dynamic and the static reconfiguration using two criteria: cost and perfo…

Improvement of Fingerprint Sensor Reading Using FPGA Devices

In order to realize fingerprint recognition system in real time environment, we describe in this paper signal controller to read fingerprint sensor generated in FPGA devices. Basically this signal is generated using state machine. The simulation result for behavioral simulation and signal generation read by logic analyzer are presented in this paper. Initialization and reading time for 76800 pixels are 50.99 mS. It is faster than fingerprint sensor using USB connection, which is more than 250 ms.

Design of Thin-Film-Transistor (TFT) arrays using current mirror circuits for Flat Panel Detectors (FPDs)

We designed 4x4 matrix TFTs arrays using current mirror amplifiers. Advantages of current mirror amplifiers are they need less requiring switches and the conversion time is short. The TFTs arrays 4x4 matrix using current mirror circuits have been fabricated and tested with success. The TFTs array directly can process signals coming from 16 pixels in the same node. This enables us to make the summation of the light intensities of close pixels during a reading.

CMOS Photodiode Design for Gamma Camera Application

We designed new photodiodes sensors including current mirror amplifiers. These photodiodes have been fabricated using a CMOS 0.6 micrometers process from Austria Micro System (AMS). The photodiode areas are respectively 1 mm × 1 mm and 0.4 mm × 0.4 mm with fill factor 98% and total chip area is 2 square millimetres. The sensor pixels show a logarithmic response in illumination and are capable of detecting very low blue light (less than 0.5 lux) . These results allow to use our sensor in new gamma camera solid-state concept.

Simple Fingerprint Minutiae Extraction Algorithm Using Crossing Number On Valley Structure

Most of the existing fingerprint extraction techniques currently available are based on ridge structure. The ridge usually has thicker structure than the valley, so that more processing time is needed to extract the ridge than extracting the valley. Taking the advantage of the thin structure of the valley, we proposed an algorithm that reduces the time needed for minutiae extraction. The algorithm was developed in Matlab environment using fingerprint images from FVC2004. In order to show the performance of the algorithm, numerical results are presented.

Image analysis of corrosion pit damage

<title>Fast motion estimation based on spatio-temporal Gabor filters: parallel implementation on multi-DSP</title>

The aim of our work is to implement a system of motion estimation in image sequences processing on DSP's: fast motion estimation based on Gabor spatio-temporal filters. Our approach consists to calculate optical flow using an energy-based method, named combined filtering which associates the energetic responses of Gabor spatio- temporal filters organized in triads. For this purpose, we applicate a technique developed by the Laboratory LIS in France, inspired from the architecture of Heeger. To reduce the computation time, we present also a parallel implementation of the algorithm on a multi-DSP architecture using SynDEx tool which is a programming environment to generate optimized distribut…

Vers une architecture électronique unifiée et zéro temps mort pour l'instrumentation nucléaire

International audience; Dans l'instrumentation nucléaire, et d'autres domaines tels que les applications de la physique des particules, l'analyse de signaux médicaux (Tomographie par Émission de Positons (TEP), Électro-EncéphaloGraphie (EEG)) ou la détection de signaux radar, le signal mesuré est composé d'impulsions dont la date d'arrivée est aléatoire, et dont l'amplitude et la durée sont non déterministes. La grande variété d'applications nécessite de disposer d'une architecture flexible pouvant être aisément reprogrammée. De plus, l'exigence de mesures en temps réel impose de disposer d'une grande capacité de calcul et de bande passante pour la mise en forme et l'extraction des caractér…

Fast Image Mosaicing for Panoramic Face Recognition

In this article, we present some development results of a system that performs mosaicing (or mosaicking) of panoramic faces. Our objective is to study the feasibility of panoramic face construction in real-time. To do so, we built a simple acquisition system composed of 5 standard cameras which, together, can take simultaneously 5 views of a face at different angles. Then, we chose an easily hardware-achievable algorithm, consisting of successive linear transformations, in order to compose a panoramic face from these 5 views. The method has been tested on a relatively large number of faces. In order to validate our system of panoramic face mosaicing, we also conducted a preliminary study on…

Procédé de codage d'un signal réel en un signal quantifié

Procédé, mis en œuvre par un processeur, de codage d'un signal réel, par exemple un signal image, en un signal quantifié, ledit procédé comprenant les étapes suivantes appliquées à chaque échantillon réel dudit signal réel : • Convertir (101) l'échantillon réel dans une représentation numérique, • Sélectionner (102), dans la partie fractionnaire du nombre, un nombre N prédéterminé de bits non nuls les plus significatifs, • pour chaque bit i significatif non nul sélectionné, i variant de 1 à N, déterminer (103) sa distance P, par rapport au bit significatif non nul sélectionné voisin de rang supérieur ou, pour le premier bit significatif non nul sélectionné, par rapport à la virgule, • retra…

Analysis of HMAX Algorithm on Black Bar Image Dataset

An accurate detection and classification of scenes and objects is essential for interacting with the world, both for living beings and for artificial systems. To reproduce this ability, which is so effective in the animal world, numerous computational models have been proposed, frequently based on bioinspired, computational structures. Among these, Hierarchical Max-pooling (HMAX) is probably one of the most important models. HMAX is a recognition model, mimicking the structures and functions of the primate visual cortex. HMAX has already proven its effectiveness and versatility. Nevertheless, its computational structure presents some criticalities, whose impact on the results has never been…

Measuring trunk orientation with a CMOS camera: feasibility and accuracy.

International audience; The purpose of this study was to develop and validate a new tool to objectively quantify trunk orientation at the bedside, especially dedicated to the measurement of the lateropulsion in acute and subacute stroke patients. We developed software to analyze 2D movement with a CMOS camera (Logitech Quickcam Pro 4000) and to calculate the orientation of a segment defined by two color markers. First, the accuracy, reproducibility and noise when measuring segment orientations were evaluated with the CMOS camera placed in different positions, and second trunk orientation was measured in static and in dynamic conditions both with a CMOS camera and with a gold standard 3D vid…

Evaluating creep in metals by grain boundary extraction using directional wavelets and mathematical morphology

Abstract It is economically important for manufacturers of high-temperature machines to be able to measure creep so that they can predict residual service life more accurately. This paper describes and refines an image analysis method for evaluating creep in laboratory test pieces. It is a preliminary study of how to extract relevant information for creep measurement by cavities counting. Sample preparation for quantification by image analysis is an important step determining the further development of the image analysis technique. Grain boundary extraction, which is directional information, is the major question to be overcome before measurement can be automated. The search for a crest-lin…

Face tracking and recognition: from algorithm to implementation

This paper describes a system capable of realizing a face detection and tracking in video sequences. In developing this system, we have used a RBF neural network to locate and categorize faces of different dimensions. The face tracker can be applied to a video communication system which allows the users to move freely in front of the camera while communicating. The system works at several stages. At first, we extract useful parameters by a low-pass filtering to compress data and we compose our codebook vectors. Then, the RBF neural network realizes a face detection and tracking on a specific board.

Toward nano-device image processing: a neuro-inspired learning approach

National audience; As nano-scale devices such as OG-CNTFETs are under studies and may be used in a near future, we choose to investigate in which application domain such components may be of the most interest. In this paper we present how neural networks can be used to implement functions on nano-scale components. This method has been tested in the image processing application field.

A 3-D simulation with virtual stereo rig for centrifugal fertilizer spreading

International audience; Stereovision can be used to characterize of the fertilizer centrifugal spreading process and to control the spreading fertilizer distribution pattern on the ground reference. Fertilizer grains, however, resemble each other and the grain images contain little information on texture. Therefore, the accuracy of stereo matching algorithms in literature cannot be used as a reference for stereo images of fertilizer grains. In order to evaluate stereo matching algorithms applied to images of grains a generator of synthetic stereo particle images is presented in this paper. The particle stereo image generator consists of two main parts: the particle 3D position generator and…

Multiscale Edges Detection by Wavelet Transform for Model of Face Recognition

Publisher Summary The linear auto-associator is a particular case of the linear-associator. The goal of this network is to associate a set of stimuli to itself, which could be used to store and retrieve face images and it also could be applied as a pre-processing device to simulate some psychological tasks—such as categorizing face according to their gender. A technique of learning based on the wavelet transform can improve recognition capability when the pattern images are with a great noise. One of the ways to store and recall face images uses the linear auto-associative memory. This connectionist model is in conjunction with a pixel-based coding of the faces. The image processing using t…

<title>Achieving high performances at lower cost for real-time image rotation by using dynamic reconfiguration</title>

FPGA components are widely used today to perform various algorithms (digital filtering) in real time. The emergence of Dynamically Reconfigurable (DR) FPGAs made it possible to reduce the number of necessary resources to carry out an image processing application (tasks chain). We present in this article an image processing application (image rotation) that exploits the FPGA's dynamic reconfiguration feature. A comparison is undertaken between the dynamic and static reconfiguration by using two criteria, cost and performance criteria. For the sake of testing the validity of our approach in terms of Algorithm and Architecture Adequacy , we realized an AT40K40 based board ARDOISE.

Real-time image rotation using B-spline interpolation on FPGA's board

The aim of our work is to realize the implementation of a real-time high-quality image rotation on FPGA's board. The method we used is based on M. Unser's work and consists in applying a B-spline interpolator. The difficulty of this problem is due to the relatively weak integration capacity of FPGAs. To solve this problem we have searched for determining the minimum number of bits to code the filter while keeping a good accuracy of filtering output. In this article, we remind a few definitions about B-spline functions and we present how we use B- spline interpolation for the image rotation problem. Then, we describe the way we calculate probability density function of the output error in or…

Development of Active Pixel Photodiode Sensors for Gamma Camera Application

We designed new photodiodes sensors including current mirror amplifiers. These photodiodes have been fabricated using a CMOS 0.6 micrometers process from Austria Micro System (AMS). The Photodiode areas are respectiveley 1mm x 1mm and 0.4mm x 0.4mm with fill factor 98 % and total chip area is 2 square millimetres. The sensor pixels show a logarithmic response in illumination and are capable of detecting very low blue light (less than 0.5 lux) . These results allow to use our sensor in new Gamma Camera solid-state concept.

Optimizing Hough transform for fertilizer spreading optical control

International audience; In Europe, centrifugal spreading is a widely used method for agricultural soil fertilization. In this broadcasting method, fertilizer particles fall onto a spinning disk, are accelerated by a vane, and afterward are ejected into the field. To predict and control the spread pattern, a low-cost, embeddable technology adapted to farm implements must be developed. We focus on obtaining the velocity and the direction of fertilizer granules when they begin their flight by means of a simple imaging system. We first show that the outlet angle of the vane is a bounded value and that its measurement provides the outlet velocity of the particle. Consequently, a simple camera un…

Implementation of JPEG2000 arithmetic decoder using dynamic reconfiguration of FPGA

This paper describes implementation of a part of JPEG2000 algorithm (MQ-Decoder and arithmetic decoder) on a FPGA board using dynamic reconfiguration. Comparison between static and dynamic reconfiguration is presented and new analysis criteria (time performance, logic cost, spatio-temporal efficiency) are defined. MQ-decoder and arithmetic decoder can be classified in the most attractive case for dynamic reconfiguration implementation: applications without parallelism by functions. This implementation is done on an architecture designed to study dynamic reconfiguration of FPGAs: the ARDOISE architecture. The implementation obtained, based on four partial configurations of arithmetic decoder…

Compression embarquée temps réel pour caméras rapides

Les caméras rapides sont de puissants outils pour étudier, par exemple, la dynamique des fluides ou le déplacement des pièces mécaniques lors d'un processus de fabrication. Nous décrivons dans ce papier, un nouveau type de caméra rapide possédant un fonctionnement original. En effet, outre le fait qu'elle utilise comme d'autres caméras, la grande flexibilité des capteurs CMOS en termes d'acquisition (ROI), elle est novatrice au niveau du transfert des données. Celles-ci pouvant être à la fois traitées et/ou compressées en temps réel au sein même de la caméra. Le transfert peut s'effectuer alors à l'aide d'une simple connection série de type USB 2.0. On réalise ainsi l'économie d'une mémoire…

<title>Restoration of a short-exposure image sequence degraded by atmospheric turbulence</title>

This paper deals with the restoration of the shape of an object observed with a high-resolution infrared imaging device, through atmospheric turbulence. The propagation path is quite long (a few tenth kilometer) and the image is thus disturbed. A sequence of short-exposure images of the interesting object is recorded. We can see that the object shape fluctuates randomly during the sequence, but that its edges remain sharp, thanks to the very short exposure time. A bayesian analysis of the Fourier descriptors associated to the edges shows that the optimal shape is the one corresponding to the mean Fourier descriptors. We thus propose two ways to estimate this shape. The first one consists in…

Fast and Robust Face Detection on a Parallel Optimized Architecture implemented on FPGA

In this paper, we present a parallel architecture for fast and robust face detection implemented on FPGA hardware. We propose the first implementation that meets both real-time requirements in an embedded context and face detection robustness within complex backgrounds. The chosen face detection method is the Convolutional Face Finder (CFF) algorithm, which consists of a pipeline of convolution and subsampling operations, followed by a multilayer perceptron. We present the design methodology of our face detection processor element (PE). This methodology was followed in order to optimize our implementation in terms of memory usage and parallelization efficiency. We then built a parallel arch…

Auto-Adaptive Trigger and Pulse Extraction for Digital Processing in Nuclear Instrumentation

International audience; This paper presents a novel auto-adaptive method for pulse triggering and extraction. Pulse triggering uses a threshold that must be placed as close as possible to the noise level. We do this by means of an adaptive threshold level based on real-time noise level estimation. A dynamic estimation of the pulse length is also used for pulse selection. The proposed approach is largely insensitive to noise and enables autonomous extraction of pulses regardless of their shape, height or length. The proposed approach can be used with numerous types of detectors from an analog-to-digital converter, and can be used in conjunction with various pulse processing techniques such a…

172 REAL TIME EDGES DETECTION USING WAVELET TRANSFORM

One of the way to extract edges uses the fast wavelet transform algorithm. This technique allows the detection of multiscale edges and is used to detect all the details, which are in a picture by modifying the scale. The real time application for edge detection involves the implementation of the algorithm on an integrated circuit like a FPGA and the development of an appropriated board. This article deals about the implementation of a wavelet transform algorithm onto a FPGA and development of an electronic board to detect multiscale edges.

Small Sample Biometric Recognition Based on Palmprint and Face Fusion

Contactless biometrics provide high comfort and hygiene in person recognition. Because of this, such systems are better accepted by the general public. This paper proposes an adaptive, contactless, biometric system which combines two modalities: palmprint and face. The processing chain has been designed to overcome embedded system constraints and small sample set problem: after a palmprint is extracted from a hand image, Gabor filters are applied to both the palmprint and face in order to extract parameters, which are then used for classification. Fusion possibilities are also discussed and tested using a multimodal database of 130 people designed by the authors. High recognition performanc…

Prefiltering for pattern recognition using wavelet transform and neural networks

Publisher Summary Neural networks are built from simple units interlinked by a set of weighted connections. Generally, these units are organized in layers. Each unit of the first layer (input layer) corresponds to a feature of a pattern that is to be analyzed. The units of the last layer (output layer) produce a decision after the propagation of information. Before feeding the computational data to neural networks, the signal must undergo a preprocessing in order to (1) define the initial transformation to represent the measured signal, (2) retain important features for class discrimination and discard that is irrelevant, and (3) reduce the volume of data to be processed, for example, data …

Automatic Evaluation of Histological Prognostic Factors Using Two Consecutive Convolutional Neural Networks on Kidney Samples

BACKGROUND AND OBJECTIVES: The prognosis of patients undergoing kidney tumor resection or kidney donation is linked to many histologic criteria. These criteria notably include glomerular density, glomerular volume, vascular luminal stenosis, and severity of interstitial fibrosis/tubular atrophy. Automated measurements through a deep-learning approach could save time and provide more precise data. This work aimed to develop a free tool to automatically obtain kidney histologic prognostic features. DESIGN, SETTING, PARTICIPANTS, & MEASUREMENTS: In total, 241 samples of healthy kidney tissue were split into three independent cohorts. The “Training” cohort (n=65) was used to train two convoluti…

Approches région et bayésienne pour la restauration ďimages dégradées par la turbulence atmosphérique

La turbulence atmospherique perturbe ľobservation a haute resolution. C’est un phenomene etudie depuis longtemps, en astronomie notamment. Le present article porte sur le cas de ľobservation ďun objet situe environ vingt kilometres, la propagation etant horizontale et pres du sol, en infrarouge. Les images a longue pose sont restaurees avec des algorithmes classiques de deconvolution. Les resultats ne sont satisfaisants que pour une faible perturbation. Il est plus avantageux ďexploiter des images courte pose, car elles contiennent plus de hautes frequences spatiales; mais ľobjet observe y fluctue aleatoirement. On travaille donc ici partir ďune sequence de plusieurs dizaines ďimages. Deux …

Parallel implementation of a multiscale edges detection algorithm

We present in this paper an implementation of a multiscale edges detection algorithm on multiprocessor using SYnDEx which is a programming environment to generate optimized distributed real-time executives. The implementation has been done on three TMS320C40 and the acceleration in comparison with one processor is 2.2.

Two-step cross correlation-based algorithm for motion estimation applied to fertilizer granules' motion during centrifugal spreading

Imaging systems are progressing in both accuracy and ro- bustness, and their use in precision agriculture is increasing accordingly. One application of imaging systems is to understand and control the cen- trifugal fertilizing spreading process. Predicting the spreading pattern on the ground relies on an estimation of the trajectories and velocities of ejected granules. The algorithms proposed to date have shown low ac- curacy, with an error rate of a few pixels. But a more accurate estimation of the motion of the granules can be achieved. Our new two-step cross- correlation-based algorithm is based on the technique used in particle image velocimetry (PIV), which has yielded highly accurate…

Design of a Real-time face detection parallel architecture using High-Level Synthesis

Abstract We describe a High-Level Synthesis implementation of a parallel architecture for face detection. The chosen face detection method is the well-known Convolutional Face Finder (CFF) algorithm, which consists of a pipeline of convolution operations. We rely on dataflow modelling of the algorithm and we use a high-level synthesis tool in order to specify the local dataflows of our Processing Element (PE), by describing in C language inter-PE communication, fine scheduling of the successive convolutions, and memory distribution and bandwidth. Using this approach, we explore several implementation alternatives in order to find a compromise between processing speed and area of the PE. We …

Système de prise d'images haute résolution pour l'analyse de projection de particules : application à l'épandage centrifuge d'engrais

This paper describes the design of a high resolution low cost imaging system for the analysis of high speed particle projection. This system, based on a camera and a set of flashes, is used to characterize the centrifugal spreading of fertilizer particles ejected at speeds of environ 30 m s. Multiexposure images collected with the camera installed perpendicular to the output flow of granules are analysed to estimate the trajectories of the fertilizer granules. Very good results are obtained with the Markov random fields method, in comparison with others.

High-speed motion estimation of fertilizer granules with Gabor filters

In the context of fertilizer supply reduction, the understanding of the whole centrifugal spreading process became essential. Since few years we focused our research on the determination by image processing of the ejection conditions of flight of the granules, that is the trajectories and ejection angles, used as input data for ballistic flight to predict the fertilizer repartition on the ground. Due to relative high speed of the fertilizer granules (around 40 m.s -1 ), the previous parameters were evaluated using a specific high speed imaging system and image processing based on motion estimation method using Markov Random Fields method (MRFs). Even if the results were good (90% of correct…

PNeuro: A scalable energy-efficient programmable hardware accelerator for neural networks

Proceedings of a meeting held 19-23 March 2018, Dresden, Germany; International audience; Artificial intelligence and especially Machine Learning recently gained a lot of interest from the industry. Indeed, new generation of neural networks built with a large number of successive computing layers enables a large amount of new applications and services implemented from smart sensors to data centers. These Deep Neural Networks (DNN) can interpret signals to recognize objects or situations to drive decision processes. However, their integration into embedded systems remains challenging due to their high computing needs. This paper presents PNeuro, a scalable energy-efficient hardware accelerat…

Cost comparison of image rotation implantations on static and dynamic Reconfigurable FPGAs

FPGA components are widely used today to perform various algorithms (digital filtering) in real time. The emergence of Dynamically Reconfigurable (DR) FPGAs made it possible to reduce the number of necessary resources to carry out an image processing application (tasks chain). We present in this article an image processing application (image rotation) that exploits the FPGA 's dynamic reconfiguration feature. A comparison is undertaken between the dynamic and static reconfiguration by using two criteria, cost and performance criteria. For the sake of testing the validity of our approach in terms of Algorithm and Architecture Adequacy, we realized an AT40K40 based board ARDOISE.

Multiple modular very long instruction word processors based on field programmable gate arrays

Modern field programmable gate array (FPGA) chips, with their large memory capacity and reconfigurability potential, are opening new frontiers in rapid prototyping of embedded systems. With the advent of high-density FPGAs, it is now possible to implement a high-performance very long instruction word (VLIW) processor core in an FPGA. This paper describes research results about enabling the DSP TMS320 C6201 model for real-time image processing applications by exploiting FPGA technology. We present a modular DSP C6201 VHDL model with a variable instruction set. We call this new development a minimum mandatory modules (M3) approach. Our goals are to keep the flexibility of DSP in order to shor…

Real-time image segmentation for anomalies detection using SVM approximation

In this paper, we propose a method of implementation improvement of the decision rule of the support vector machine, applied to real-time image segmentation. We present very high speed decisions (approximately 10 ns per pixel) which can be useful for detection of anomalies on manufactured parts. We propose an original combination of classifiers allowing fast and robust classification applied to image segmentation. The SVM is used during a first step, pre-processing the training set and thus rejecting any ambiguities. The hyperrectangles-based learning algorithm is applied using the SVM classified training set. We show that the hyperrectangle method imitates the SVM method in terms of perfor…

A pre-processing technique based on the wavelet transform for linear autoassociators with applications to face recognition

In order to improve the performance of a linear autoassociator (which is a neural network model), we explore the use of several preprocessing techniques. The gist of our approach is to store, in addition to the original pattern, one or several pre-processed (i.e. filtered) versions of the patterns to be stored in a neural network. First, we compare the performance of several pre-processing techniques (a plain vanilla version of the autoassociator as a control, a Sobel operator, a Canny-Deriche operator, and a multiscale Canny-Deriche operator) on an example of a pattern completion task using a noise degraded version of a face stored in an autoassociator. We found that the multiscale Canny-D…

Real-time flaw detection on complex part: Study of SVM and hyperrectangle based method

We present in this paper the study of two classifications methods used in order to control in real-time some industrials parts. We present the practical frame in which is made the operations, natures of the anomaly to be detected as well as the features extractions method. We tested two techniques of classification, with different algorithm complexities and performances. We compare the results obtained on various features spaces. We end by a combinatorial perspective of results of classification.

Multiple Modular VLIW Processors based on FPGA

International audience

Event-Based Trajectory Prediction Using Spiking Neural Networks

International audience; In recent years, event-based sensors have been combined with spiking neural networks (SNNs) to create a new generation of bio-inspired artificial vision systems. These systems can process spatio-temporal data in real time, and are highly energy efficient. In this study, we used a new hybrid event-based camera in conjunction with a multi-layer spiking neural network trained with a spike-timing-dependent plasticity learning rule. We showed that neurons learn from repeated and correlated spatio-temporal patterns in an unsupervised way and become selective to motion features, such as direction and speed. This motion selectivity can then be used to predict ball trajectory…

Parallel implementation on DSPs of a face detection algorithm

In order to localize the face in an image, our approach consists of approximating the face oval shape with an ellipse and to compute coordinates of the center of the ellipse. For this purpose, we explore a new version of the Hough transformation: the fuzzy generalized Hough transformation. To reduce the computation time, we present also a parallel implementation of the algorithm on 2 digital signal processors and we show that an acceleration of a factor of 1.62 has been obtained.

<title>Real-time face tracking and recognition for video conferencing</title>

This paper describes a system of vision in real time, allowing to detect automatically the faces presence, to localize and to follow them in video sequences. We verify also the faces identities. These processes are based by combining technique of image processing and methods of neural networks. The tracking is realized with a strategy of prediction-verification using the dynamic information of the detection. The system has been evaluated quantitatively on 8 video sequences. The robustness of the method has been tested on various lightings images. We present also the analysis of complexity of this algorithm in order to realize an implementation in real time on a FPGA based architecture.

A Parallel Face Detection System Implemented on FPGA

In this paper, we introduce a methodology for designing a system for face detection and its implementation on FPGA. The chosen face detection method is the well-known convolutional face finder (CFF) algorithm, which consists in a pipeline of convolutions and subsampling operations. Our goal is to define a parallel architecture able to process efficiently this algorithm. We present a dataflow based architecture algorithm adequation (AAA) methodology implemented using the SynDEx software, in order to find the best compromise between the processing power and functionality requirement of each processor element (PE), and the efficiency of algorithm parallelization. We describe a first implementa…

An Enhanced Detector of Blurred and Noisy Edges

Detecting edges in digital images is a tricky operation in image processing since images may contain areas with different degrees of noise, blurring and sharpness. Such operation represents an important step of the whole process of similarity shape analysis and retrieval.

Real-time Sub-pixel Cross Bar Position Metrology

Many measurement application fields need to calculate cross bar intersection locations of horizontal and vertical bars. The system we developed and that we present in this paper is an embedded system that measures cross bar positions with sub-pixel accuracy on 1024×1024 pixel images delivered by a camera at a 50 MHz data rate in real time. This is done using an algorithm that looks for intersection areas and then locally calculates two lines representing horizontal and vertical bars. The two line intersection is considered to be the bar intersection. To achieve real time, we developed a hybrid architecture in which low level processes are implemented into FPGAs and others into DSPs. As a re…

Implementation of pattern recognition algorithm based on RBF neural network

In this paper, we present implementations of a pattern recognition algorithm which uses a RBF (Radial Basis Function) neural network. Our aim is to elaborate a quite efficient system which realizes real time faces tracking and identity verification in natural video sequences. Hardware implementations have been realized on an embedded system developed by our laboratory. This system is based on a DSP (Digital Signal Processor) TMS320C6x. The optimization of implementations allow us to obtain a processing speed of 4.8 images (240x320 pixels) per second with a correct rate of 95% of faces tracking and identity verification.