Search

The Two-Dimensional Fast Fourier Transform (2D-FFT) algorithm is used for the study of many modern systems applied for security and biometrics. The adoption of this algorithm, which is a compute intensive task, is limited due to its hardware design complexity. The first objective of this paper is to underline the effect of the hardware/software co-design (Hw/Sw co-design) for the reduction of the processing time and power consumption. Secondly, we propose an innovative architecture for the 2D-FFT algorithm tested on Zynq Soc, which requires less processing time and memory compared to the traditional algorithm. Three implementations (one software and two Hw/Sw co-designs) of the 2D-FFT algorithm using the Zynq SoC are presented in this paper. The first is based on ARM processor. A speedup of 29x is obtained compared to the original implementation thanks to many optimizations. The second is a Hw/Sw co-design solution of the traditional 2D-FFT algorithm introduced on a hybrid platform combining an ARM Cortex-A9 processor with an FPGA. The third is also a Hw/Sw co-design solution using our optimized 2D-FFT algorithm to reach the real-time contraints for high-resolution images (1920 × 1080). It provides a speedup of 1.13x, 3.31x and 96.21x faster than the Hw/Sw co-design implementation of the traditional RC algorithm, the pure software implementations with and without optimizations, respectively.

International audience; The brute-force behaviour of High-Efficiency Video Coding (HEVC) is the biggest hurdle in the communication of the multimedia contents. Therefore, a novel method will be presented here to expedite the intra mode decision process of HEVC. In the first step, the feasibility of the odds-algorithm for the early intra mode decision is presented by using statistical evidence. Then, various elements of odds algorithm are analyzed and then mapped to the intra mode process (elements) of HEVC. Finally, the probability required by the odds algorithm is obtained by utilizing the correlation between the current and the neighboring blocks. The proposed algorithm accelerated the encoding process of the HEVC by 25% to 35%, while the Bjontegaard Delta Bit Rate (BD-BR) is 0.95% to 1.84%.

International audience; High-Efficiency Video Coding (HEVC) applies 35 intra modes to every block of a frame and selects the mode that gives the best prediction. This brute-force nature of HEVC makes it complex and unfit for real-time applications. Therefore, a fast intra-mode estimation algorithm is presented here based on the classic World War II (WW2) technique known as the ‘German Tanks Problem’ (GTP). This not only is the first article to use GTP for early estimation of intra mode, but also expedites the estimation process of GTP. Secondly, the various elements of the intra process are efficiently mapped to the elements of GTP estimation. Finally, the two variations of GPT are modeled and are also minimum-variance estimates. These experimental results indicate that proposed GTP-based fast estimation reduced the compression time of HEVC from 23.88% to 31.44%.

In several particularly secure applications such as the entrance to a school, it is important to know whether the person entering is an adult or a child. In this article, we propose a human body morphology detector that distinguishes whether the person is an adult or a child. This detector could be included in a smart portal to detect whether the entry person is an adult or a child to apply a different treatment depending on the morphology. A person detector module1 is deployed to detect the presence of a person within a predefined radius. When the location of the person is detected, our system can measure the height of the person and determine if the person is an adult or a child based on its height.

Behavioral analysis in an urban environment is a complex task that requires material and human resources, due to the difficulty of interpreting the situations. This paper presents a method to improve the detection of dangerous behaviors by assisting surveillance stations. Our objective is to alert when one of these behaviors is captured by a surveillance camera. To do this, we analyze the positions and paths of the persons in a global way, through a group of parameters. These parameters are determined by an automatic image analysis algorithm such as DBSCAN computed on an NVIDIA Jetson TX2. This analysis allows to detect, through the evolution and clustering of points in each cloud, phenomena qualified as abnormal, such as dispersion and rapid clustering, as well as poaching. The data used to feed our algorithm come from simulations that allow testing new and different scenarios. The performance of our proposed method is evaluated on videos representing real case situations.

International audience; In this study, we propose a novel method to encrypt fully-phase information combining the concepts of the transport of intensity equation and spatially variant polarization encoding. The transport of intensity equation is a non-iterative and non-interferometric phase-retrieval method which recovers the phase information from defocused intensities. Spatially variant polarization encoding employs defocused intensity measurements. The proposed cryptosystem uses a two-step optical experimentation process—primarily, a simple set-up for defocused intensities recording for phase retrieval and then a set-up for encoding. Strong security, convenient intensity-based measurements, and noise-free decryption are the main features of the proposed method. The simulation results have been presented in support of the proposed idea. However, the TIE section of the cryptosystem, as of now, has been experimentally demonstrated for micro-lens.

In this article, we present a new method of dehazing based on the Koschmieder model, which aims to restore an image that has been affected by haze. The difficulty is to improve the estimation of the transmission and the atmospheric light that generally suffer from the nonhomogeneity and the random variability of the environment. The keypoint is to enhance the dehazing of very bright regions of the image in order to improve the treatment of the sky that is often overestimated or underestimated compared to the rest of the scene. The approach proposed in this paper is based on two main contributions: 1. an L0 gradient optimization function weighted by a set of Gaussian filters and based on an iterative algorithm for optimization convergence. Unlike the existing methods using a single value of the atmospheric light for the whole image, our method uses a set of values neighboring an initial estimated value. The fusion is then applied based on Laplacian and Gaussian pyramids to combine all the relevant information from the set of images constructed from atmospheric lights and improves the contrast to recover the colors of the sky without any artifacts. Finally, the results are validated by three criteria: an autocorrelation score (ZNCC), a similarity measure (SSIM) and a visual criterion. The experiments carried out on two datasets show that our approach allows a better dehazing of the images with higher SSIM and ZNCC measurements but also with better visual quality.

International audience; Pattern detection is an active field in big data streams analytics with numerous ongoing challenges. Actually, due to the great velocity and variety of data, new patterns can appear and change over time. Existing state-of-the-art solutions consist in updating the pattern detection model regularly in order to integrate newly appeared and validated patterns. However, in several applications, such as security and defense, patterns can represent anomalies. Therefore, it becomes crucial to detect new patterns (i.e. new anomalies), as early as possible, in order to react at the right moment. Consequently, emergent pattern detection becomes a very challenging task. To tackle this challenge, we propose EPDA (Emergent Pattern Detection Algorithm): a new and validated algorithm for detecting emergent patterns in data streams. The originality of EPDA consists in exploiting frequent pattern mining techniques by proposing new statistical measures in order to estimate the evolution of emergent patterns over time. To perform this detection in a real-time, EPDA runs on the well-known Apache STORM distributed real-time computation system. To better fit our algorithm, we propose a new Apache STORM topology which is composed of one Spouts level and two Bolts levels. Experiments on a real data stream have shown the relevance of the proposed measures and the efficiency of our algorithm in a prediction task and in terms of execution time.

International audience; In order to accompany the swimming coaches in evaluating high-level swimmers, we developed a prototype for instantaneous speed estimation. To achieve this, we proposed and validated, in a previous work, a swimmer tracking system based on data fusion. However, the initialization phase is done manually, and our aim, in this paper, is to automate this process. First, we propose a region of interest localization module that allows the detection of the first appearance of the swimmer in the lane as well as the restriction of the region of interest around him. This module is based on the method a contrario which consists of modeling the random noise corresponding to the water and detecting the structured movement relative to the swimmer motion. To do that, we calibrate the pool using DLT (Direct Linear Transform) technique, extract the concerned lane, apply the frame difference approach to detect the moving objects, and then decompose the lane into blocs and classify them into swimmer motion or noise. Second, in order to detect the swimmer’s head, we propose the Scaled Composite JTC which is based on the NL-JTC correlation technique. The input plane of this latter includes a target and a reference image. The first is the region of interest detected by the method a contrario. The second consists of a Scaled Composite Reference. The tests conducted on real video sequences of French swimming championships (Limoges 2015) showed very good results in terms of region of interest localization and swimmer’s head detection which allows a reliable initialization for the tracking system.

Optical correlation has a rich history in image recognition applications from a database. In practice, it is simple to implement optically using two lenses or numerically using two Fourier transforms. Even if correlation is a reliable method for image recognition, it may jeopardize decision making according to the location, height, and shape of the correlation peak within the correlation plane. Additionally, correlation is very sensitive to image rotation and scale. To overcome these issues, in this study, we propose a method of nonparametric modelling of the correlation plane. Our method is based on a kernel estimation of the regression function used to classify the individual images in the correlation plane. The basic idea is to improve the decision by taking into consideration the energy shape and distribution in the correlation plane. The method relies on the calculation of the Hausdorff distance between the target correlation plane (of the image to recognize) and the correlation planes obtained from the database (the correlation planes computed from the database images). Our method is tested for a face recognition application using the Pointing Head Pose Image Database (PHPID) database. Overall, the results demonstrate good performances of this method compared to competitive methods in terms of good detection and very low false alarm rates.

International audience; This study investigates the combination of Mueller imaging polarimetry with machine learning for the automated optical classication of raw materials. It shows that standard image classication techniques based on support vector machines or deep neural networks can readily be applied to polarimetric data extracted from Mueller matrix measurements. The feasability of such an approach is empirically demonstrated through the classication of multispectral depolarization images of real-world materials (banana, wood and foam samples).

International audience; Object identification in highly turbid optical media depends mainly on the quality of collected images. Underwater images acquired in a turbid environment are generally of very poor quality. Attenuation and backscattering of light by water, by materials dissolved in the water, and by particulate material are the main causes of the degradation of underwater images. It is therefore essential to improve the quality of such images to facilitate object identification. The focus of this paper is to report the principle and validation of a fast and effective method of improving the quality of underwater images. On the one hand, this method uses a polarimetric imaging optical system to reduce the effect of diffusion on the image acquisition. On the other hand, it is based on an optimized version of the dark channel prior (DCP) method that has received a great deal of attention for image dehazing. Results derived from images obtained in a controlled laboratory water tank environment with different turbidity conditions and images from tests using the proposed method at sea demonstrate an ability to significantly improve visibility and reduce runtime by a factor of about 50 for a 4K image when compared to conventional DCP methods.

This paper describes a new approximate transform for the high efficiency video coding (HEVC). A 8 × 8 discrete cosine transform (DCT) approximation is proposed and then down-sampled or expanded to generate the 4 × 4, 16 × 16, and 32 × 32 approximate matrices. The proposed 8 × 8 approximation is carried out in part by neighbourhood in order to take the advantage of adjacent pixels correlation of natural images. Hence, rather than approximating the odd basis vectors of DCT kernel by referring to their intrinsic values, we choose to quantize that by taking into account their signs and positions. The proposed approximation matrices respect the properties of transform matrices prescribed by HEVC like orthogonality and bit-length of the basis vector elements. Furthermore, they have nearly the same arithmetic complexity and hardware requirement as those of recently proposed related methods, but involve significantly less error energy. Moreover, a reconfigurable design based on the 8 × 8 approximation transform is proposed in order to allow the simultaneous computation of eight 4-, four 8-, two 16-, or one 32-point approximate DCTs. It is found that the reconfigurable design can involve nearly 26% less area-delay product (ADP) when compared with the separate non-reconfigurable designs. Experimental results obtained from FPGA prototype and HM simulations have demonstrated the advantages of the proposed transforms.

International audience; This study is concerned with the design of a Mueller imaging polarimeter for the visualization of spatially varying Mueller matrix fields. A simplified calibration procedure is advocated, where all the optical elements are calibrated simultaneously rather than independently as in the state of the art. This is shown to significantly reduce the bias inherent to sequential calibration methods. In addition, this procedure requires no reference sample, it allows calibration both in transmission or in reflection modes, and it relies on ready-to-use cameras. Put together, these novelties should help nonspecialists in optics in designing and calibrating a Mueller imaging polarimeter for applications such as material classification.

International audience; Approximation of discrete cosine transform (DCT) is useful for reducing its computational complexity without significant impact on its coding performance. Most of the existing algorithms for approximation of the DCT target only the DCT of small transform lengths, and some of them are non-orthogonal. This paper presents a generalized recursive algorithm to obtain orthogonal approximation of DCT where an approximate DCT of length N could be derived from a pair of DCTs of length (N/2) at the cost of N additions for input preprocessing. We perform recursive sparse matrix decomposition and make use of the symmetries of DCT basis vectors for deriving the proposed approximation algorithm. Proposed algorithm is highly scalable for hardware as well as software implementation of DCT of higher lengths, and it can make use of the existing approximation of 8-point DCT to obtain approximate DCT of any power of two length, N > 8. We demonstrate that the proposed approximation of DCT provides comparable or better image and video compression performance than the existing approximation methods. It is shown that proposed algorithm involves lower arithmetic complexity compared with the other existing approximation algorithms. We have presented a fully scalable reconfigurable parallel architecture for the computation of approximate DCT based on the proposed algorithm. One uniquely interesting feature of the proposed design is that it could be configured for the computation of a 32-point DCT or for parallel computation of two 16-point DCTs or four 8-point DCTs with a marginal control overhead. The proposed architecture is found to offer many advantages in terms of hardware complexity, regularity and modularity. Experimental results obtained from FPGA implementation show the advantage of the proposed method.

International audience; Techniques are widely sought to detect and identify sea mines. This issue is characterized by complicated mine shapes and underwater light propagation dependencies. In a preliminary study we use a preprocessing step for denoising underwater images before applying the algorithm for mine detection. Once a mine is detected, the protocol for identifying it is activated. Among many correlation filters, we have focused our attention on the asymmetric segmented phase-only filter for quantifying the recognition rate because it allows us to significantly increase the number of reference images in the fabrication of this filter. Yet they are not entirely satisfactory in terms of recognition rate and the obtained images revealed to be of low quality. In this report, we propose a way to improve upon this preliminary study by using a single wavelength polarimetric camera in order to denoise the images. This permits us to enhance images and improve depth visibility. We present illustrative results using in situ polarization imaging of a target through a milk-water mixture and demonstrate that our challenging objective of increasing the detection rate and decreasing the false alarm rate has been achieved.

International audience; This study presents and validates an optimized method of simultaneous compression and encryption designed to process images with close spectra. This approach is well adapted to the compression and encryption of images of a time-varying scene but also to static polarimetric images. We use the recently developed spectral fusion method [Opt. Lett. 35, 1914–1916 (2010)] to deal with the close resemblance of the images. The spectral plane (containing the information to send and/or to store) is decomposed in several independent areas which are assigned according a specific way. In addition, each spectrum is shifted in order to minimize their overlap. The dual purpose of these operations is to optimize the spectral plane allowing us to keep the low- and high-frequency information (compression) and to introduce an additional noise for reconstructing the images (encryption). Our results show that not only can the control of the spectral plane enhance the number of spectra to be merged, but also that a compromise between the compression rate and the quality of the reconstructed images can be tuned. We use a root-mean-square (RMS) optimization criterion to treat compression. Image encryption is realized at different security levels. Firstly, we add a specific encryption level which is related to the different areas of the spectral plane, and then, we make use of several random phase keys. An in-depth analysis at the spectral fusion methodology is done in order to find a good trade-off between the compression rate and the quality of the reconstructed images. Our new proposal spectral shift allows us to minimize the image overlap. We further analyze the influence of the spectral shift on the reconstructed image quality and compression rate. The performance of the multiple-image optical compression and encryption method is verified by analyzing several video sequences and polarimetric images.

An automated surveillance method that uses multiple image processing can detect, analyze, and track movements to identify emergency situations.

International audience; Multiple input multiple output (MIMO) technology with equal gain-combining (EGC) is Considered in this work to Enhance the data rate of free-space optical communication (FSO) system in log-normal channel. In FSO system, Propagating light in free space is used to transmit data for communication system. 256-ary pulse position modulation (PPM) using intensity modulation with Direct Detection (IM / DD) as a power efficient system is employed using PIN receiver. Obtained results Demonstrate That significant boost of data rate can be Achieved using MIMO techniques. Enhances data rate from 350 Mbps to 800 Mbps When using 2 × 2 MIMO configuration. As well, is reported about 1 Gbps for 4 × 4 MIMO system. These enhancements are Achieved without any bandwidth or power expansions and the bit error rate (BER) at the receiver is maintained below 10 ^-3.

One promising solution to reduce the computational complexity of discrete cosine transform (DCT) is to identify the redundant computations and to get rid of them. In this study, the authors present a new method to predict zero-quantised DCT coefficients for efficient implementation of intra-frame video encoding by identifying such redundant computations. Traditional methods use the Gaussian statistical model of residual pixels to predict all-zero or partial-zero blocks. The proposed method is based on two key ideas. At first, the bounds of DCT coefficients are derived from the intermediate signals of the Loeffler DCT algorithm instead of calculating the sum of absolute difference (SAD) of residual pixels. The sufficiency conditions are then suitably chosen to predict the zero-quantised coefficients to reduce the arithmetic complexity without degrading the video quality. Simulation results are found to validate the analytical model and show that the proposed prediction eliminates more redundant computations than the existing methods. Moreover, the authors have derived a pipelined VLSI architecture of the proposed prediction scheme which offers a saving of more than 63 and 91% of multiplications of the second stage of one-dimensional DCT for high and low bit-rate intra-video encoding, respectively.

International audience; In this paper, we defined a low complexity 2D-DCT architecture. The latter will be able to transform spatial pixels to spectral pixels while taking into account the constraints of the considered compression standard. Indeed, this work is our first attempt to obtain one reconfigurable multistandard DCT. Due to our new matrix decomposition, we could define one common 2D-DCT architecture. The constant multipliers can be configured to handle the case of RealDCT and/or IntDCT (multiplication by 2). Our optimized algorithm not only provides a reduction of computational complexity, but also leads to scalable pipelined design in systolic arrays. Indeed, the 8 × 8 StdDCT can be computed by using 4×4 StdDCT which can be obtained by calculating 2×2 StdDCT. Besides, the proposed structure can be extended to deal with higher number of N (i.e. 16 × 16 and 32 × 32). The performance of the proposed architecture are better when compared with conventional designs. In particular, for N = 4, it is found that the proposed design have nearly third the area-time complexity of the existing DCT structures. This gain is expected to be higher for a greater size of 2D-DCT.

International audience; Underwater target detection is investigated by combining active polarization imaging and optical correlation-based approaches. Experiments were conducted in a glass tank filled with tap water with diluted milk or seawater and containing targets of arbitrary polarimetric responses. We found that target estimation obtained by imaging with two orthogonal polarization states always improves detection performances when correlation is used as detection criterion. This experimentally study illustrates the potential of polarization imaging for underwater target detection and opens interesting perspectives for the development of underwater imaging systems.

We introduce a double optimization procedure for spectrally multiplexing multiple images. This technique is adapted from a recently proposed optical setup implementing the discrete cosine transformation (DCT). The new analysis technique is a combination of spectral fusion based on the properties of DCT, specific spectral filtering, and quantization of the remaining encoded frequencies using an optimal number of bits. Spectrally multiplexing multiple images defines a first level of encryption. A second level of encryption based on a real key image is used to reinforce encryption. A set of numerical simulations and a comparison with the well known JPEG (Joint Photographic Experts Group) image compression standard have been carried out to demonstrate the improved performances of this method. The focus here will differ from the method of simultaneous fusion, compression, and encryption of multiple images (SFCE) [Opt. Express 19, 24023 (2011)] in the following ways. Firstly, we shall be concerned with optimizing the compression rate by adapting the size of the spectral block to each target image and decreasing the number of bits required to encode each block. This size adaptation is achieved by means of the root-mean-square (RMS) time-frequency criterion. We found that this size adaptation provides a good tradeoff between bandwidth of spectral plane and number of reconstructed output images. Secondly, the encryption rate is improved by using a real biometric key and randomly changing the rotation angle of each block before spectral fusion. By using a real-valued key image we have been able to increase the compression rate of 50% over the original SFCE method. We provide numerical examples of the effects for size, rotation, and shifting of DCT-blocks which play noteworthy roles in the optimization of the bandwidth of the spectral plane. Inspection of the results for different types of attack demonstrates the robustness of our procedure.

International audience; In this paper we present a novel architecture for FFT implementation on FPGA. The proposed architecture based on radix-4 algorithm presents the advantage of a higher throughput and low area-delay product. In fact, the novelty consists on using a memory sharing and dividing technique along with parallel-in parallel-out Processing Elements (PE). The proposed architecture can perform N-point FFT using only 4/3N delay ele- ments and involves a latency of N/4 cycles. Comparison in terms of hardware complexity and area-delay product with recent works presented in the literature and commercial IPs has been made to show the efficiency of the proposed design. Moreover, from the experimental results obtained from a FPGA prototype we find that the proposed design involves an execution time of 56% lower than that obtained with Xilinx IP core and an increase of 19% in the throughput by area ratio for 256-point FFT.

International audience; An optimized technique based on the fringe-adjustedJTC (joint transform correlator) architecture is proposed and validated for rotation invariant recognition and tracking of a target in an unknown input scene. To enhance the robustness of the proposed technique, we used a three-step optimization by: (1) utilizing the fringe-adjusted filter (HFAF) in the Fourier plane, (2) adding nonlinear processing in the Fourier plane, and (3) using a new decision criterion in the correlation plane by considering the correlation peak energy and the highest peaks outside the desired correlation peak. Several tests were conducted to reduce the number of reference images needed for fast tracking while ensuring robust discrimination and efficient tracking of thedesired target. Test results obtained using the PHPID (Pointing Head Pose Image Database) data baseconfirm robust performance of the proposed method for face recognition and tracking applications. Thereafter, we also tested the proposed technique for a challenging application i.e. underwater mine detection and excellent results were obtained.

International audience; Interestingly, the past 20 years have provided us many examples of optical correlation methods for pattern recognition, e.g. VanderLugt correlator (VLC). In recent years, hybrid techniques, i.e. numerical implementation of correlation, have been also considered an alternative to all-optical methods because they show a good compromise between performance and simplicity. Moreover, these correlation methods can be implemented using an all-numerical and reprogrammable target such as the graphics processor unit (GPU), or the field-programmable gate array (FPGA). However, this numerical procedure requires realizing two Fourier Transforms (FT), a spectral multiplication, and a correlation plane analysis. The purpose of this study is to compare the performances of a numerical correlator based on the fast Fourier transform (FFT) with that relying on a simulation of the Fraunhofer diffraction. Different tests using the Pointing Head Pose Image Database (PHPID) and considering faces with vertical and horizontal rotations were performed with the code MATLAB. Tests were conducted with a five reference optimized composite filter. The receiving operating characteristics (ROC) curves show that the optical FT simulating the Fraunhofer diffraction leads to better performances than the FFT. The implications of our results for correlation are discussed.

International audience; Optical correlation is based on the comparison of contours between an unknown target image and a known reference image. However, it does not usually include the color image information in the decision making process. In order to render the correlation method sensitive to color change, we propose a general method based on the decomposition of the target image in its three color components using, either the normalized RGB (red, green, blue) color space, or the normalized HSV (hue, saturation, value) space. Then, the correlation operation is carried out for each color component and the results are merged in order to make a decision. The aforementioned steps can alleviate some of the problems associated with illumination changes in the target image but do utilize color information of the target image. To overcome these problems, we propose to convert the color information in contour information into a signature corresponding to the color information of the target image. This technique and test results are presented to validate its effectiveness. The preliminary results obtained with this technique are encouraging.

International audience; We suggest a new type of optimized composite filter, i.e. the asymmetric segmented phase-only filter (ASPOF), for improving the effectiveness of a VanderLugt correlator (VLC) when used for face identification. Basically, it consists in merging several reference images after application of a specific spectral optimization method. After segmentation of the spectral filter plane to several areas, each area is assigned to a single winner reference according to a new optimized criterion. The point of the paper is to show that this method offers a significant performance improvement on standard composite filters for face identification. We first briefly revisit composite filters (adapted, phase-only, inverse, compromise optimal, segmented, minimum average correlation energy, optimal trade-off maximum average correlation, and amplitude modulated phase-only (AMPOF)) which are tools of choice for face recognition based on correlation techniques and compare their performances with those of the ASPOF. We illustrate some of the drawbacks of current filters for several binary and gray scale images identifications. Next, we describe the optimization steps and introduce the ASPOF that can overcome these technical issues to improve the quality and the reliability of the correlation based decision. We derive performance measures, i.e. peak-to-correlation energy values and receiver operating characteristic curves, to confirm consistency of the results. We numerically find that this filter increases the recognition rate and decrease the false alarm rate. The results show that the discrimination of the ASPOF is comparable to that of the AMPOF, but the ASPOF is more robust than the trade-off maximum average correlation height (OT-MACH) against rotation and various types of noise sources. Our method has several features that make it amenable to experimental implementation using a VLC.

International audience; The key outcome of this work is to propose and validate a fast and robust correlation scheme for face recognition applications. The robustness of this fast correlator is ensured by an adapted pre-processing step for the target image allowing us to minimize the impact of its (possibly noisy and varying) amplitude spectrum information. A segmented composite filter is optimized, at the very outset of its fabrication, by weighting each reference with a specific coefficient which is proportional to the occurrence probability. A hierarchical classification procedure (called two-level decision tree learning approach) is also used in order to speed up the recognition procedure. Experimental results validating our approach are obtained with a prototype based on GPU implementation of the all-numerical correlator using the NVIDIA GPU GeForce 8400 GS processor and test samples from the Pointing Head Pose Image Database (PHPID), e.g. true recognition rates larger than 85 % with a run time lower than 120 ms have been obtained using fixed images from the PHPID, true recognition rates larger than 77 % using a real video sequence with 2 frame per second when the database contains 100 persons. Besides, it has been shown experimentally that the use of more recent GPU processor like NVIDIA-GPU Quadro FX 770M can perform the recognition of 4 frame per second with the same length of database

International audience; The Discrete Cosine Transform (DCT)-based image com- pression is widely used in today's communication systems. Signi cant research devoted to this domain has demonstrated that the optical com- pression methods can o er a higher speed but su er from bad image quality and a growing complexity. To meet the challenges of higher im- age quality and high speed processing, in this chapter, we present a joint system for DCT-based image compression by combining a VLSI archi- tecture of the DCT algorithm and an e cient quantization technique. Our approach is, rstly, based on a new granularity method in order to take advantage of the adjacent pixel correlation of the input blocks and to improve the visual quality of the reconstructed image. Second, a new architecture based on the Canonical Signed Digit and a novel Common Subexpression Elimination technique is proposed to replace the constant multipliers. Finally, a recon gurable quantization method is presented to e ectively save the computational complexity. Experimental results obtained with a prototype based on FPGA implementation and com- parisons with existing works corroborate the validity of the proposed optimizations in terms of power reduction, speed increase, silicon area saving and PSNR improvement.

International audience; We report a new spectral multiple image fusion analysis based on the discrete cosine transform (DCT) and a specific spectral filtering method. In order to decrease the size of the multiplexed file, we suggest a procedure of compression which is based on an adapted spectral quantization. Each frequency is encoded with an optimized number of bits according its importance and its position in the DC domain. This fusion and compression scheme constitutes a first level of encryption. A supplementary level of encryption is realized by making use of biometric information. We consider several implementations of this analysis by experimenting with sequences of gray scale images. To quantify the performance of our method we calculate the MSE (mean squared error) and the PSNR (peak signal to noise ratio). Our results consistently improve performances compared to the wellknown JPEG image compression standard and provide a viable solution for simultaneous compression and encryption of multiple images.

International audience; In this paper, our objective is to detail know-how and techniques that can help the designer of electronic circuits to develop and to optimize their own IP in a reasonable time. For this reason, we propose to optimize existing FFT algorithms for low-cost FPGA implementations. For that, we have used short length structures to obtain higher length transforms. Indeed, we can obtain a VLSI structure by using log4 (N) 4-point FFTs to construct N-point FFT rather than (N/8) log8 (N) 8-point FFTs. Furthermore, two techniques are used to yield with VLSI architecture. Firstly, the radix-4 FFT is modified to process one sample per clock cycle. Secondly, the memory is shared and divided into 4 parts to reduce the consumed resources and to improve the overall latency. Comparisons with commercial IP cores show that the low area architecture presents the best compromise in terms of speed/area.

International audience; Home automation is being implemented into more and more domiciles of the elderly and disabled in order to maintain their independence and safety. For that purpose, we propose and validate a surveillance video system, which detects various posture-based events. One of the novel points of this system is to use adapted Vander-Lugt correlator (VLC) and joint-transfer correlator (JTC) techniques to make decisions on the identity of a patient and his three-dimensional (3-D) positions in order to overcome the problem of crowd environment.We propose a fuzzy logic technique to get decisions on the subject's behavior. Our system is focused on the goals of accuracy, convenience, and cost, which in addition does not require any devices attached to the subject. The system permits one to study and model subject responses to behavioral change intervention because several levels of alarm can be incorporated according different situations considered. Our algorithm performs a fast 3-D recovery of the subject's head position by locating eyes within the face image and involves a model-based prediction and optical correlation techniques to guide the tracking procedure. The object detection is based on (hue, saturation, value) color space. The system also involves an adapted fuzzy logic control algorithm to make a decision based on information given to the system. Furthermore, the principles described here are applicable to a very wide range of situations and robust enough to be implementable in ongoing experiments

International audience; In this paper, we propose a new technique for rotation invariant recognition and tracking of the face of a target person in a given scene. We propose an optimized method for face tracking based on the Fringe-adjusted JTC architecture. To validate our approach, we used the PHPID data base containing faces with various in-plane rotations. To enhance the robustness of the proposed method, we used a three-step optimization technique by: (1) utilizing the fringe-adjusted filter (HFAF) in the Fourier plane, (2) adding nonlinearity in the Fourier plane after applying the HFAF filter, and (3) using a new decision criterion in the correlation plane by considering the correlation peak energy and five largest peaks outside the highest correlation peak. Several tests were made to reduce the number of reference images needed for fast tracking while ensuring robust discrimination and efficient of the desired target.

International audience; In this paper, we explore the use of optical correlation-based recognition to identify and position underwater man-made objects (e.g. mines). Correlation techniques can be defi ned as a simple comparison between an observed image (image to recognize) and a reference image; they can be achieved extremely fast. The result of this comparison is a more or less intense correlation peak, depending on the resemblance degree between the observed image and a reference image coming from a database. However, to perform a good correlation decision, we should compare our observed image with a huge database of references, covering all the appearances of objects we search. Introducing all the appearances of objects can influence speed and/or recognition quality. To overcome this limitation, we propose to use composite filter techniques, which allow the fusion of several references and drastically reduce the number of needed comparisons to identify observed images. These recent techniques have not yet been exploited in the underwater context. In addition, they allow for integrating some preprocessing directly in the correlation filter manufacturing step to enhance the visibility of objects. Applying all the preprocessing in one step reduces the processing by avoiding unnecessary Fourier transforms and their inverse operation. We want to obtain fi lters that are independent from all noises and contrast problems found in underwater videos. To achieve this and to create a database containing all scales and viewpoints, we use as references 3D computer-generated images.

This chapter covers some of the latest developments in optical correlation techniques for face recognition using the concept of spectral fusion and placing a special emphasis on its practical aspects and applications for face recognition. Optical correlation is reviewed as a method to carry out instantly a decision on the target form (form to be recognized). A range of relevant practical topics is discussed, such as JTC (Joint Transform Correlator) and the Vander-Lugt architectures. Both of them are based on the "4f" optical setup and on a comparison between the target image and a reference image. The similarity between the two images is achieved by the detection of a correlation peak. The development of suitable algorithms for optical correlation and new technologies in electro-optical interfaces has significantly improved the processing capacity of optical correlators and reduce their sizes. To overcome the limitations of a decision taken by a simple detection of a correlation peak, more complex decision algorithms are required. These algorithms necessitate the integration of new correlation filters and the realization of multiple correlations and reconfigurable multi-channel architectures.

We propose and analyze a dual encryption/decryption scheme, motivated by recent interest in polarization encoding. Compared to standard optical encryption methods, which are based on phase and amplitude manipulation, this encryption procedure relying on Mueller-Stokes formalism provides large flexibility in the key encryption design. The effectiveness of our algorithm is discussed, thanks to a numerical simulation of the polarization encryption/decryption procedure of a 256 gray-level image. Of additional special interest is the immunity of this encryption algorithm to brute force attacks.

International audience; In this document we are presenting a new approach to design an optimised Direct Digital Frequency Synthesizer (DDFS) for complex demodulation used in digital receivers. For that, we suggest an adaptation of the phase to sine converter by combining the two following techniques: 1) an optimized COordinate Rotation DIgital Computer (CORDIC) algorithm 2) the principle of Taylor series approximation. To validate our proposed approach, a DDFS with 8 Hz tuning frequency resolution and 20 bits output data (for sine and cosine waves) is being implemented in Xilinx FPGA device giving a maximum operating frequency of more than 306 MHz and a Spurious Free Dynamic Range (SFDR) of 112 dBc. To prove the good performances of the proposed approach, we compared it favorably with several existing DDFS architectures.

International audience; In some recognition form applications (which require multiple images: facial identification or sign-language), many images should be transmitted or stored. This requires the use of communication systems with a good security level (encryption) and an acceptable transmission rate (compression rate). In the literature, several encryption and compression techniques can be found. In order to use optical correlation, encryption and compression techniques cannot be deployed independently and in a cascade manner. Otherwise, our system will suffer from two major problems. In fact, we cannot simply use these techniques in a cascade manner without considering the impact of one technique over another. Secondly, a standard compression can affect the correlation decision, because the correlation is sensitive to the loss of information. To solve both problems, we developed a new technique to simultaneously compress & encrypt multiple images using a BPOF optimized filter. The main idea of our approach consists in multiplexing the spectrums of different transformed images by a Discrete Cosine Transform (DCT). To this end, the spectral plane should be divided into several areas and each of them corresponds to the spectrum of one image. On the other hand, Encryption is achieved using the multiplexing, a specific rotation functions, biometric encryption keys and random phase keys. A random phase key is widely used in optical encryption approaches. Finally, many simulations have been conducted. Obtained results corroborate the good performance of our approach. We should also mention that the recording of the multiplexed and encrypted spectra is optimized using an adapted quantification technique to improve the overall compression rate.

International audience; Over the years extensive studies have been carried out to apply coherent optics methods in real-time communications and image transmission. This is especially true when a large amount of information needs to be processed, e.g., in high-resolution imaging. The recent progress in data-processing networks and communication systems has considerably increased the capacity of information exchange. However, the transmitted data can be intercepted by nonauthorized people. This explains why considerable effort is being devoted at the current time to data encryption and secure transmission. In addition, only a small part of the overall information is really useful for many applications. Consequently, applications can tolerate information compression that requires important processing when the transmission bit rate is taken into account. To enable efficient and secure information exchange, it is often necessary to reduce the amount of transmitted information. In this context, much work has been undertaken using the principle of coherent optics filtering for selecting relevant information and encrypting it. Compression and encryption operations are often carried out separately, although they are strongly related and can influence each other. Optical processing methodologies, based on filtering, are described that are applicable to transmission and/or data storage. Finally, the advantages and limitations of a set of optical compression and encryption methods are discussed.

International audience; Abstract--Biometric security systems such as Retina scanner, fingerprint and face recognition among others are the most recently developed security systems. The latest systems can be found in various situations (police checkpoints, airport facilities, . . . ). As they become widely used and requested, an increasing number of researchers are working in this field. In this paper, we propose a new Biometric Recognition method based on Independent Component Analysis technique "ICA".

International audience; In the last two decades, wireless communications have been introduced in various applications. However, the transmitted data can be, at any moment, intercepted by non-authorized person. That could be explain why data encryption and secure transmission have gained enormous popularity. In order to secure data transmission, one should pay attention to two problems: transmission rate and encryption security level. In this manuscript, we address the last two problems by proposing a new videoimage transmission scheme. The new system consists in using the advantage of optical high transmission rate and some powerful signal processing tools to secure the transmitted data. The main idea of our approach is to secure the transmitted information at two levels: classical level using an adaptation of standard optical technic and second level (spatial diversity) using independent transmitters. In the second level, a hacker should intercept not one channel but all of them in order to retrieve the information. At designed receiver, one can easily apply ICA algorithms to decrypt the received signals and retrieve the information.