This paper deals with the topic of the retrieval of document images focused on a specific application: the ornaments of the Hand-Press period. It presents an overview as a result of the work and the discussions undertaken by a workgroup on this subject. The paper starts by giving a general view about digital libraries of ornaments and associated retrieval problematics. Two main issues are underlined: content based image retrieval (CBIR) and image difference visualization. Several contributions are summarized, commented and compared. Conclusions and open problems arising from this overview are twofold: 1. contributions on CBIR miss scale-invariant methods and don't provide significative evaluation results. 2. robust registration is the open problem for visual comparison.

In this paper, an original method for texture comparison and classification is presented. It is based on an adaptation of the Gray Local Dissimilarity Map (GLDM) for the comparison of textured images. In our method, GLDM of gray level co-occurrence matrices (GLCM) are computed instead of GLDM of images directly. Only one parameter is extracted from these GLDMs and used to classify textures. The method is tested on a texture dataset with two-class and multi-class classification using K-Nearest Neighbours (KNN) to prove its efficiency. The obtained results show that computing the GLDM of GLCM gives better performance for texture classification than both computing the GLDM of images and classification methods based on the extraction of several characteristics from GLCM.

Due to the considerable increase of images in everyday life, many applications require a study on their similarity. The main challenge is to find a simple and efficient method to compare and classify image pairs into similar and dissimilar classes. This study presents a new method to image pairs comparison and classification based on the modeling of the Local Dissimilarity Map (LDM). The LDM is a tool for locally measuring the dissimilarity between two binary or grayscale images. It is a measure of dissimilarities based on a modified version of the Hausdorff distance, which allows quantifying locally the dissimilarities between images. This measure is completely without parameters and generic. The image pairs classification (2-class classification) method is structured as follows. First, a statistical model for the LDM is proposed. The model parameters, used as descriptors, are relevant to discriminate similar and dissimilar image pairs. Second, classifiers are applied to compute the classification scores (2-class classification problem). In addition, this approach is robust with respect to geometric transformations such as translation compared to the state-of-the-art similarity measures. Although the main objective of this paper is to apply our approach to image pairs classification, it is also performed on a classification with more than two classes (multi-class classification). Experiments on the well-known image data sets ^∗NIST and on old print data set prove that the proposed method produces comparable, even better results than the state-of-the-art methods in terms of accuracy and F(1) score.

In previous works, the Local Dissimilarity Map (LDM) was proposed to compare two binary and grayscale images. This measure is based on a Hausdorff distance, which allows to quantify locally the dissimilarities between images. In this paper, we proposed the two-parameter Weibull distribution to model the LDM and the undersampled LDMs for two structural images. To classify structural image pairs, we used the two parameters of Weibull distribution for each LDM as descriptors. They are relevant to discriminate image pairs into similar and dissimilar classes. Experiments were made on the MNIST image dataset and in our own old print image dataset. The results shown our approach is more accurate than the other measures in the literature.

Considering the unavailability of labeled data sets in remote sensing change detection, this letter presents a novel and low complexity unsupervised change detection method based on the combination of similarity and dissimilarity measures: mutual information (MI), disjoint information (DI), and local dissimilarity map (LDM). MI and DI are calculated on sliding windows with a step of 1 pixel for each pair of channels of both images. The resulting scalar values, weighted by q and m coefficients, are multiplied by the values of the center pixels of the windows weighted by p to remove the textures on images. The changes are detected using, respectively, the grayscale LDM and color LDM. A sliding window is then used on the color LDM and each pixel is characterized by a two-parameter Weibull distribution. Binarized change maps can be obtained by using a k-means clustering on the model parameters. Experiments on optical aerial image data set show that the proposed method produces comparable, even better results, to the state-of-the-art methods in terms of recall, precision, and F-measure.

This paper presents a new change detection technique for images taken from the sentinel-2 satellite between 2015 and 2018 in different regions of the world. These images are widely used in recent years for change detection. This technique is based on two dissimilarity measures: the Disjoint Information and the Local Dissimilarity Map. The disjoint information quantifies the dissimilarities between textures and the Local Dissimilarity Map those between structures of images. Firstly, the disjoint information is computed across the blocks of the RGB image channels and the value is multiplied by the center value of the pixel of each block. Secondly, the Local Dissimilarity Maps over the pre-processed channels and the average of the pixel values on the Local Dissimilarity Maps are computed. Finally, an extension of the Gaussian OTSU's threshold is used to detect changes in images. Experimental results on the well-known Onera Satellite Change Detection (OSCD) dataset show the effectiveness of our proposed method compared to the state-of-the-art deep learning methods.

This work aims at exploring a general framework embedding techniques from classifiers, Time–Frequency Distributions (TFD) and dissimilarity measures for epileptic seizures detection. The proposed approach consists firstly in computing dissimilarities between TFD of electroencephalogram (EEG) signals and secondly in using them to define a decision rule. Compared to the existing approaches, the proposed one uses entire TFD of EEG signals and does not require arbitrary feature extraction. Several dissimilarity measures and TFDs have been compared to select the most appropriate for EEG signals. Classifiers, such as Artificial Neural Network (ANN), Support Vector Machine (SVM), Linear Discriminate Analysis (LDA) and k-Nearest Neighbours (k-NN), have been combined with the proposed approach. In order to evaluate the proposed approach, 13 different classification problems (including 2, 3 and 5-class) pertaining to five types of EEG signals have been used. The comparison between results obtained with the proposed approach and results reported in the literature with the same database of epileptic EEG signals demonstrates the effectiveness of this approach for seizure detection. Experimental results show that this approach has achieved highest accuracy in the most studied classification problems. A high value of 98% is achieved for the 5-class problem. Further, in most classification problems with 2 and 3-class, it also yields a satisfactory accuracy of approximately 100%. The robustness of the proposed approach is evaluated with the addition of noise to the EEG signals at various signal-to-noise ratios (SNRs). The experimental results show that this approach has a good classification accuracy at low SNRs.

Because of the overlapping of the spectra of different metabolites and the interference of the baseline mainly from broad resonances of macromolecule and lipids, it is difficult to achieve the quantification of spectra of different metabolites which is important for both research and clinical applications of Magnetic Resonance Spectroscopy (MRS). In this paper, a novel MRS quantification method based on frequency a priori knowledge is proposed. Firstly, a wavelet filter is used to remove the broad components of an observed spectrum in which baseline and the relatively broad components of metabolite spectrum are included. Secondly, a linear nonnegative pursuit algorithm based on regularized FOCUSS (Focal Underdetermined System Solver) algorithm is used to decompose the filtered spectra in a dictionary which is based on a set of Lorentzian and Gaussian functions corresponding to spectrum models. Benefitting from the a priori knowledge of the peak frequency of each metabolite, the filtered metabolite spectrum can be sparsely represented with these basis functions and the spectra of different metabolites are relevant to certain basis functions. Therefore, with the corresponding relation between the basis functions and spectrum models and the estimated decomposition coefficients, a mixed spectrum without baseline can be reconstructed and spectra of different metabolites can be quantified at the same time. The accuracy and the robustness of MRS quantification are improved by the proposed method, from simulation data, compared with commonly used MRS quantification methods. Quantification on in vivo brain spectra is also demonstrated.

In this paper, a sparse representation method is proposed for magnetic resonance spectroscopy (MRS) quantification. An observed MR spectrum is composed of a set of metabolic spectra of interest, a baseline and a noise. To separate the spectra of interest, the a priori knowledge about these spectra, such as signal models, the peak frequencies, and linewidth ranges of different resonances, is first integrated to construct a dictionary. The separation of the spectra of interest is then performed by using a pursuit algorithm to find their sparse representations with respect to the dictionary. For the challenging baseline problem, a wavelet filter is proposed to filter the smooth and broad components of both the observed spectra and the basis functions in the dictionary. The computation of sparse representation can then be carried out by using the remaining data. Simulation results show the good performance of this wavelet filtering-based strategy in separating the overlapping components between the baselines and the spectra of interest, when no appropriate model function for the baseline is available. Quantifications of in vivo brain MR spectra from tumor patients in different stages of progression demonstrate the effectiveness of the proposed method.

This paper proposes a frequency-domain Magnetic Resonance (MR) spectral processing method based on sparse representation for accurate quantification of prostate spectra. Generally, an observed prostate spectrum can be considered as a mixture of resonances of interest, a baseline and noise. As the resonances of interest often overlap and the baseline is unknown, their separation and quantification can be difficult. In the proposed method, based on the commonly used signal model of prostate spectra and some a priori knowledge of nonlinear model parameters, a dictionary is constructed which can sparsely represent the resonances of interest as well as the baseline in an input spectrum. The estimation of the resonances of interest is achieved by finding their sparse representations with respect this dictionary. A linear pursuit algorithm based on regularized FOCUSS (Focal Underdetermined System Solver) algorithm is proposed to estimate these sparse representations. The robustness and accuracy of prostate spectrum quantification of the proposed method are improved compared with two classical spectral processing methods: model-based time domain fitting and frequency-domain analysis based on peak integration when tested on simulation data. Quantification on in vivo prostate spectra is also demonstrated and the results appear encouraging.

Accurate prostate cancer localization with noninvasive imaging can be used to guide biopsy, radiotherapy and surgery as well as to monitor disease progression [1]. Published studies have shown multispectral magnetic resonance imaging (MRI), i.e., a combination of multiple types of MR images, as a promising noninvasive imaging technique for the localization of prostate cancer.

This paper presents a tumor detecting system that allows interactive 3D tumor visualization and tumor volume measurements. An improved level set method is proposed to automatically segment the tumor images slice by slice. PET images are used to detect the tumor while CT images make a 3D representation of the patient's body possible. An initial slice with a seed within the tumor is firstly chosen by the operator. The system then performs automatically the tumor volume segmentation that allows the clinician to visualize the tumor, to measure it and to evaluate the best medical treatment adapted to the patient.

This article describes a new method for ancient books ornamental letters segmentation and recognition. The purpose of our work is to automatically determine the letter represented in an ornamental letter image. Our process is divided in two parts: a segmentation step of the ornamental letter is followed by a recognition step. The segmentation process uses multiresolution analysis to filter background decorations followed by a binarisation step and a morphologic reconstruction of the expected letter. The recognition process use the previously obtained reconstruction and compares it with capital letters images used as a dictionary of shapes with the Local Dissimilarity Map (LDM) distance.

In this article a research work in the field of content-based multiresolution indexing and retrieval of images is presented. Our method uses multiresolution decomposition of images using wavelets - in the HSV colorspace - to extract parameters at multiple scales allowing a progressive (coarse-to-fine) retrieval process. Features are automatically classified into several clusters with K-means algorithm. A model image is computed for each cluster in order to represent all the images of this cluster, The process is reiterated again and again and each cluster is sub-divided into sub-clusters. The model images are stored in a tree which is proposed to users for browsing the database. The nodes of the tree are the families and the leaves are the images of the database. A paleontology images database is used to test the proposed technique. This kind of approach permits to build a visual interface easy to use for users. Our main contribution is the building of the tree with multiresolution indexing and retrieval of images and the generation of model images to be proposed to users.

This article describes a content-based indexing and retrieval (CBIR) system based on hierarchical binary signatures. Binary signatures are obtained through a described binarization process of classical features (color, texture and shape). The Hamming binary distance (based on binary XOR operation) is used during retrieval. This technique was tested on a real image collection containing 7200 images and on a virtual collection of one million images. Results are very good both in terms of speed and accuracy allowing real-time image retrieval in very large image collections.

This article proposes a content-based indexing and retrieval (CBIR) system based on query-by-visual-example using hierarchical binary signatures. Binary signatures are obtained through a described binarization process of classical features (color, texture and shape). The Hamming binary distance (based on binary XOR operation) is used for computing distances. This technique was tested on a real natural image collection containing 10 000 images and on a virtual collection of one million images. Results are very good both in terms of speed and accuracy allowing near real-time image retrieval in very large image collections.

This article presents a visual browsing content-based indexing and retrieval (CBIR) system for large image databases applied to a paleontology database. The studied system offers a hierarchical organization of feature vectors into signature vectors leading to a research tree so that users can explore the database visually. To build the tree, our technique consists in transforming the images using multiresolution analysis in order to extract features at multiple scales. Then a hierarchical signature vector for each scale is built using extracted features. An automatic classification of the obtained signatures is performed using the k-means algorithm. The images are grouped into clusters and for each cluster a model image is computed. This model image is inserted into a research tree proposed to users to browse the database visually. The process is reiterated and each cluster is split into sub-clusters with one model image per cluster, giving the nodes of the tree. The multiresolution approach combined with the organized signature vectors offers a coarse-to-fine research during the retrieval, process (i.e. during the progression in the research tree).

This paper presents optimized signal and image processing libraries from Intel Corporation. Intel Performance Primitives (IPP) is a low-level signal and image processing library developed by Intel Corporation to optimize code on Intel processors. Open Computer Vision library (OpenCV) is a high-level library dedicated to computer vision tasks. This article describes the use of both libraries to build flexible and efficient signal and image processing applications.

Visual browsing is an important way of searching for images in large databases. In image retrieval, a lot of problems have to be solved to get a good system: dimensionality curse, users' search context, size of the database, visual features. In this article, a method trying to attenuate these problems is proposed. Each features vector is organized into four signature vectors used in the classification process while building a fuzzy search tree that is proposed to users for visual browsing. Our system gives good results in terms of speed and accuracy by solving several problems of classical image retrieval methods.

In this article a research work in the field of content-based image retrieval in large database applied to the Paleontolgy image database of the université de Bourgogne, Dijon, France called "TRANS'TYFIPAL" is proposed. Our indexing method is based on multiresolution decomposition of database images using wavelets. For each kind of paleontology images we try to find a characteristic image representing it. This model image is computed using a classification algorithm on the space of parameters extracted from the wavelet transform of each image. Then a search tree is built to offer users a graphic interface for retrieving images. So that users have to navigate through this tree to find an image similar that of to their request. Our contribution in the field is the building of the model and of the search tree to make user access easier and faster. This paper ends with a conclusion on first coming results and a description of future work to be done to enhance our indexing and retrieval method.

In this article we present research work in the field of content-based image retrieval in large databases applied to the paleontology image database of the Université de Bourgogne, Dijon, France, called "TRANS'TYFIPAL". Our indexing method is based on multiresolution decomposition of database images using wavelets. For each family of paleontology images we try to find a model image that represents it. The K-means automatic classification algorithm divides the space of parameters into several clusters. A model image for each cluster is computed from the wavelet transform of each image of the cluster. Then a search tree is built to offer users a graphic interface for retrieving images. So users have to navigate through this tree of model images to find an image similar to that they are requesting. Our contribution in the field is the building of the model and of the search tree to make user access easier and faster. At the end of this article we give experimental results and a description of future work that will be done to enhance our indexing and retrieval method.

This communication deals with finding the position of a reference shape in a given image. The proposed matcher is constructed from local dissimilarity maps. These maps allow to efficiently and robustly measure the differences between two images. It is shown an example that the matcher potentially returns less false-positives than a reference method (chamfer matching). This is possible as the local dissimilarity measure is symmetric, which makes it more robust to noise.We show that the proposed matcher is a generalization of the chamfer matching. It also allows fast computation times. A good robustness to noise is confirmed from presented simulations.

In order to evaluate performance quality of coding techniques, it is needed to have a good global index and a local index allowing the localisation of the distortions. In this study, a local dissimilarity map is presented for gray-level images. Its application to the comparison of a compressed image and its reference allows an excellent visual detection of the distortions. A global dissimilarity index is computed from the local dissimilarity map. These new measures are compared to the structural similarity index (SSIM). The results of the global measure are as good as the SSIM. The results of the local measure are quite superior to the SSIM computed in a local window. We claim these good results come from the consistency of the proposed index. It is more consistent to compute a global measure from a local one, than a local measure from a global one.

Current presence-sensing technologies for energy-efficient lighting control and building optimisation are (i) catered to commercial and institutional environments, and (ii) focused on lamp technology and occupancy detection. They often ignore user behaviour characteristics, which significantly influence energy consumption. Therefore, this study aims to identify alternative sensing techniques as part of a lighting control system that can energy-efficiently support user's behavioural needs in mixed-function residential spaces. An exploratory study investigated the optimal placement of a non-wearable radar sensor to detect an imitated user's breathing frequency at varying pre-set horizontal distance positions, and the sensor's performance was validated with a spirometer. The procedure measured a balloon's radar-detected distance, radar-detected breathing frequency, and spirometer-registered breathing frequency at each pre-set position. The radar sensor detected all simulated breathing frequencies with almost 100% data accuracy but was not comparable in detecting all distances. The radar offers a less intrusive short-range presence-sensing for homes, accurately detecting breathing frequencies in a contactless way between 0.2m to 0.8m. Further investigations are recommended to develop radar sensing that could predict lighting options based on user's objective feedback.