arXiv subjects

Michel Salomon

Publications and source records attributed to Michel Salomon.

14 recordsLinked to original sources

arXiv2021

Deep Learning methods for automatic evaluation of delayed enhancement-MRI. The results of the EMIDEC challenge

A key factor for assessing the state of the heart after myocardial infarction (MI) is to measure whether the myocardium segment is viable after reperfusion or revascularization therapy. Delayed enhancement-MRI or DE-MRI, which is performed several minutes after injection of the contrast agent, provides high contrast between viable and nonviable myocardium and is therefore a method of choice to evaluate the extent of MI. To automatically assess myocardial status, the results of the EMIDEC challenge that focused on this task are presented in this paper. The challenge's main objectives were twofold. First, to evaluate if deep learning methods can distinguish between normal and pathological cases. Second, to automatically calculate the extent of myocardial infarction. The publicly available database consists of 150 exams divided into 50 cases with normal MRI after injection of a contrast agent and 100 cases with myocardial infarction (and then with a hyperenhanced area on DE-MRI), whatever their inclusion in the cardiac emergency department. Along with MRI, clinical characteristics are also provided. The obtained results issued from several works show that the automatic classification of an exam is a reachable task (the best method providing an accuracy of 0.92), and the automatic segmentation of the myocardium is possible. However, the segmentation of the diseased area needs to be improved, mainly due to the small size of these areas and the lack of contrast with the surrounding structures.

eess.IV↗

arXiv2017

On the ability to reconstruct ancestral genomes from Mycobacterium genus

Technical signs of progress during the last decades has led to a situation in which the accumulation of genome sequence data is increasingly fast and cheap. The huge amount of molecular data available nowadays can help addressing new and essential questions in Evolution. However, reconstructing evolution of DNA sequences requires models, algorithms, statistical and computational methods of ever increasing complexity. Since most dramatic genomic changes are caused by genome rearrangements (gene duplications, gain/loss events), it becomes crucial to understand their mechanisms and reconstruct ancestors of the given genomes. This problem is known to be NP-complete even in the "simplest" case of three genomes. Heuristic algorithms are usually executed to provide approximations of the exact solution. We state that, even if the ancestral reconstruction problem is NP-hard in theory, its exact resolution is feasible in various situations, encompassing organelles and some bacteria. Such accurate reconstruction, which identifies too some highly homoplasic mutations whose ancestral status is undecidable, will be initiated in this work-in-progress, to reconstruct ancestral genomes of two Mycobacterium pathogenetic bacterias. By mixing automatic reconstruction of obvious situations with human interventions on signaled problematic cases, we will indicate that it should be possible to achieve a concrete, complete, and really accurate reconstruction of lineages of the Mycobacterium tuberculosis complex. Thus, it is possible to investigate how these genomes have evolved from their last common ancestors.

q-bio.GN↗

arXiv2017

Improving Blind Steganalysis in Spatial Domain using a Criterion to Choose the Appropriate Steganalyzer between CNN and SRM+EC

Conventional state-of-the-art image steganalysis approaches usually consist of a classifier trained with features provided by rich image models. As both features extraction and classification steps are perfectly embodied in the deep learning architecture called Convolutional Neural Network (CNN), different studies have tried to design a CNN-based steganalyzer. The network designed by Xu et al. is the first competitive CNN with the combination Spatial Rich Models (SRM) and Ensemble Classifier (EC) providing detection performances of the same order. In this work we propose a criterion to choose either the CNN or the SRM+EC method for a given input image. Our approach is studied with three different steganographic spatial domain algorithms: S-UNIWARD, MiPOD, and HILL, using the Tensorflow computing platform, and exhibits detection capabilities better than each method alone. Furthermore, as SRM+EC and the CNN are both only trained with a single embedding algorithm, namely MiPOD, the proposed method can be seen as an approach for blind steganalysis. In blind detection, error rates are respectively of 16% for S-UNIWARD, 16% for MiPOD, and 17% for HILL on the BOSSBase with a payload of 0.4 bpp. For 0.1 bpp, the respective corresponding error rates are of 39%, 38%, and 41%, and are always better than the ones provided by SRM+EC.

cs.MM↗

arXiv2016

Relation between Gene Content and Taxonomy in Chloroplasts

The aim of this study is to investigate the relation that can be found between the phylogeny of a large set of complete chloroplast genomes, and the evolution of gene content inside these sequences. Core and pan genomes have been computed on \textit{de novo} annotation of these 845 genomes, the former being used for producing well-supported phylogenetic tree while the latter provides information regarding the evolution of gene contents over time. It details too the specificity of some branches of the tree, when specificity is obtained on accessory genes. After having detailed the material and methods, we emphasize some remarkable relation between well-known events of the chloroplast history, like endosymbiosis, and the evolution of gene contents over the phylogenetic tree.

q-bio.GN↗

arXiv2016

Binary Particle Swarm Optimization versus Hybrid Genetic Algorithm for Inferring Well Supported Phylogenetic Trees

The amount of completely sequenced chloroplast genomes increases rapidly every day, leading to the possibility to build large-scale phylogenetic trees of plant species. Considering a subset of close plant species defined according to their chloroplasts, the phylogenetic tree that can be inferred by their core genes is not necessarily well supported, due to the possible occurrence of problematic genes (i.e., homoplasy, incomplete lineage sorting, horizontal gene transfers, etc.) which may blur the phylogenetic signal. However, a trustworthy phylogenetic tree can still be obtained provided such a number of blurring genes is reduced. The problem is thus to determine the largest subset of core genes that produces the best-supported tree. To discard problematic genes and due to the overwhelming number of possible combinations, this article focuses on how to extract the largest subset of sequences in order to obtain the most supported species tree. Due to computational complexity, a distributed Binary Particle Swarm Optimization (BPSO) is proposed in sequential and distributed fashions. Obtained results from both versions of the BPSO are compared with those computed using an hybrid approach embedding both genetic algorithms and statistical tests. The proposal has been applied to different cases of plant families, leading to encouraging results for these families.

cs.AI↗

arXiv2016

Neural Networks and Chaos: Construction, Evaluation of Chaotic Networks, and Prediction of Chaos with Multilayer Feedforward Networks

Many research works deal with chaotic neural networks for various fields of application. Unfortunately, up to now these networks are usually claimed to be chaotic without any mathematical proof. The purpose of this paper is to establish, based on a rigorous theoretical framework, an equivalence between chaotic iterations according to Devaney and a particular class of neural networks. On the one hand we show how to build such a network, on the other hand we provide a method to check if a neural network is a chaotic one. Finally, the ability of classical feedforward multilayer perceptrons to learn sets of data obtained from a dynamical system is regarded. Various Boolean functions are iterated on finite states. Iterations of some of them are proven to be chaotic as it is defined by Devaney. In that context, important differences occur in the training process, establishing with various neural networks that chaotic behaviors are far more difficult to learn.

cs.NE↗

arXiv2016

Protein Folding in the 2D Hydrophobic-Hydrophilic (HP) Square Lattice Model is Chaotic

Among the unsolved problems in computational biology, protein folding is one of the most interesting challenges. To study this folding, tools like neural networks and genetic algorithms have received a lot of attention, mainly due to the NP-completeness of the folding process. The background idea that has given rise to the use of these algorithms is obviously that the folding process is predictable. However, this important assumption is disputable as chaotic properties of such a process have been recently highlighted. In this paper, which is an extension of a former work accepted to the 2011 International Joint Conference on Neural Networks (IJCNN11), the topological behavior of a well-known dynamical system used for protein folding prediction is evaluated. It is mathematically established that the folding dynamics in the 2D hydrophobic-hydrophilic (HP) square lattice model, simply called "the 2D model" in this document, is indeed a chaotic dynamical system as defined by Devaney. Furthermore, the chaotic behavior of this model is qualitatively and quantitatively deepened, by studying other mathematical properties of disorder, namely: the indecomposability, instability, strong transitivity, and constants of expansivity and sensitivity. Some consequences for both biological paradigms and structure prediction using this model are then discussed. In particular, it is shown that some neural networks seems to be unable to predict the evolution of this model with accuracy, due to its complex behavior.

q-bio.BM↗

arXiv2016

Steganalysis via a Convolutional Neural Network using Large Convolution Filters for Embedding Process with Same Stego Key

For the past few years, in the race between image steganography and steganalysis, deep learning has emerged as a very promising alternative to steganalyzer approaches based on rich image models combined with ensemble classifiers. A key knowledge of image steganalyzer, which combines relevant image features and innovative classification procedures, can be deduced by a deep learning approach called Convolutional Neural Networks (CNN). These kind of deep learning networks is so well-suited for classification tasks based on the detection of variations in 2D shapes that it is the state-of-the-art in many image recognition problems. In this article, we design a CNN-based steganalyzer for images obtained by applying steganography with a unique embedding key. This one is quite different from the previous study of {\em Qian et al.} and its successor, namely {\em Pibre et al.} The proposed architecture embeds less convolutions, with much larger filters in the final convolutional layer, and is more general: it is able to deal with larger images and lower payloads. For the "same embedding key" scenario, our proposal outperforms all other steganalyzers, in particular the existing CNN-based ones, and defeats many state-of-the-art image steganography schemes.

cs.MM↗

arXiv2015

Improved Core Genes Prediction for Constructing well-supported Phylogenetic Trees in large sets of Plant Species

The way to infer well-supported phylogenetic trees that precisely reflect the evolutionary process is a challenging task that completely depends on the way the related core genes have been found. In previous computational biology studies, many similarity based algorithms, mainly dependent on calculating sequence alignment matrices, have been proposed to find them. In these kinds of approaches, a significantly high similarity score between two coding sequences extracted from a given annotation tool means that one has the same genes. In a previous work article, we presented a quality test approach (QTA) that improves the core genes quality by combining two annotation tools (namely NCBI, a partially human-curated database, and DOGMA, an efficient annotation algorithm for chloroplasts). This method takes the advantages from both sequence similarity and gene features to guarantee that the core genome contains correct and well-clustered coding sequences (\emph{i.e.}, genes). We then show in this article how useful are such well-defined core genes for biomolecular phylogenetic reconstructions, by investigating various subsets of core genes at various family or genus levels, leading to subtrees with strong bootstraps that are finally merged in a well-supported supertree.

q-bio.GN↗

arXiv2015

Hybrid Genetic Algorithm and Lasso Test Approach for Inferring Well Supported Phylogenetic Trees based on Subsets of Chloroplastic Core Genes

The amount of completely sequenced chloroplast genomes increases rapidly every day, leading to the possibility to build large scale phylogenetic trees of plant species. Considering a subset of close plant species defined according to their chloroplasts, the phylogenetic tree that can be inferred by their core genes is not necessarily well supported, due to the possible occurrence of "problematic" genes (i.e., homoplasy, incomplete lineage sorting, horizontal gene transfers, etc.) which may blur phylogenetic signal. However, a trustworthy phylogenetic tree can still be obtained if the number of problematic genes is low, the problem being to determine the largest subset of core genes that produces the best supported tree. To discard problematic genes and due to the overwhelming number of possible combinations, we propose an hybrid approach that embeds both genetic algorithms and statistical tests. Given a set of organisms, the result is a pipeline of many stages for the production of well supported phylogenetic trees. The proposal has been applied to different cases of plant families, leading to encouraging results for these families.

cs.AI↗

arXiv2014

Gene Similarity-based Approaches for Determining Core-Genes of Chloroplasts

In computational biology and bioinformatics, the manner to understand evolution processes within various related organisms paid a lot of attention these last decades. However, accurate methodologies are still needed to discover genes content evolution. In a previous work, two novel approaches based on sequence similarities and genes features have been proposed. More precisely, we proposed to use genes names, sequence similarities, or both, insured either from NCBI or from DOGMA annotation tools. Dogma has the advantage to be an up-to-date accurate automatic tool specifically designed for chloroplasts, whereas NCBI possesses high quality human curated genes (together with wrongly annotated ones). The key idea of the former proposal was to take the best from these two tools. However, the first proposal was limited by name variations and spelling errors on the NCBI side, leading to core trees of low quality. In this paper, these flaws are fixed by improving the comparison of NCBI and DOGMA results, and by relaxing constraints on gene names while adding a stage of post-validation on gene sequences. The two stages of similarity measures, on names and sequences, are thus proposed for sequence clustering. This improves results that can be obtained using either NCBI or DOGMA alone. Results obtained with this quality control test are further investigated and compared with previously released ones, on both computational and biological aspects, considering a set of 99 chloroplastic genomes.

cs.NE↗

arXiv2014

Finding the Core-Genes of Chloroplasts

Due to the recent evolution of sequencing techniques, the number of available genomes is rising steadily, leading to the possibility to make large scale genomic comparison between sets of close species. An interesting question to answer is: what is the common functionality genes of a collection of species, or conversely, to determine what is specific to a given species when compared to other ones belonging in the same genus, family, etc. Investigating such problem means to find both core and pan genomes of a collection of species, \textit{i.e.}, genes in common to all the species vs. the set of all genes in all species under consideration. However, obtaining trustworthy core and pan genomes is not an easy task, leading to a large amount of computation, and requiring a rigorous methodology. Surprisingly, as far as we know, this methodology in finding core and pan genomes has not really been deeply investigated. This research work tries to fill this gap by focusing only on chloroplastic genomes, whose reasonable sizes allow a deep study. To achieve this goal, a collection of 99 chloroplasts are considered in this article. Two methodologies have been investigated, respectively based on sequence similarities and genes names taken from annotation tools. The obtained results will finally be evaluated in terms of biological relevance.

cs.CE↗

arXiv2011

Building a Chaotic Proved Neural Network

Chaotic neural networks have received a great deal of attention these last years. In this paper we establish a precise correspondence between the so-called chaotic iterations and a particular class of artificial neural networks: global recurrent multi-layer perceptrons. We show formally that it is possible to make these iterations behave chaotically, as defined by Devaney, and thus we obtain the first neural networks proven chaotic. Several neural networks with different architectures are trained to exhibit a chaotical behavior.

cs.AI↗

arXiv2009

Java Technology : a Strategic Solution for Interactive Distributed Applications

In a world demanding the best performance from financial investments, distributed applications occupy the first place among the proposed solutions. This particularity is due to their distributed architecture which is able to acheives high performance. Currently, many research works aim to develop tools that facilitate the implementation of such applications. The urgent need for such applications in all areas pushes researchers to accelerate this process. However, the lack of standardization results in the absence of strategic decisions taken by computer science community. In this article, we argue that Java technology represents an elegant compromise ahead of the list of the currently available solutions. In fact, by promoting the independence of hardware and software, Java technology makes it possible to overcome pitfalls that are inherent to the creation of distributed applications.

cs.DC↗