Emilie Poisson Caillault

No reference and reduced reference video quality metrics for end to end QoS monitoring

This paper describes an objective measurement method designed to assess the perceived quality for digital videos. The proposed approach can be used either in the context of a reduced reference quality assessment or in the more challenging situation where no reference is available. In that way, it can be deployed in a QoS monitoring strategy in order to control the end-user perceived quality. The originality of the approach relies on the very limited computation resources which are involved, such a system could be integrated quite easily in a real time application. It uses a convolutional neural network (CNN) that allows a continuous time scoring of the video. Experiments conducted on different MPEG-2 videos, with bit rates ranging from 2 to 6 Mbits/s, show the effectiveness of the proposed approach. More specifically, a linear correlation criterion, between objective and subjective scoring, ranging from 0.90 up to 0.95 has been obtained on a set of typical TV videos in the case of a reduced reference assessment. Without any reference to the original video, the correlation criteria remains quite satisfying since it still lies between 0.85 and 0.90, which is quite high with respect to the difficulty of the task, and equivalent and more in some cases than the traditional PSNR, which is a full reference measurement.

MS-TDNN with global discriminant trainings

This article analyses the behavior of various hybrid architectures based on a multi-state neuro-Markovian scheme (MS-TDNN HMM) applied to online handwriting word recognition systems. We have considered different cost functions, including maximal mutual information criteria with discriminant training and maximum likelihood estimation, to train the systems globally at the word level and also we varied the number of states from one up to three to model the basic hidden Markov models at the letter level. We report experimental results for non-constrained, writer independent, word recognition obtained on the IRONOFF database.

MIXED DISCRIMINANT TRAINING OF HYBRID ANN/HMM SYSTEMS FOR ONLINE HANDWRITTEN WORD RECOGNITION

Online handwritten word recognition systems usually rely on Hidden Markov Models (HMMs), which are effective under many circumstances, but suffer some major limitations in real world applications. Artificial neural networks (ANN) appear to be a promising alternative, however they failed to model sequence data such as online handwriting due to their variable lengths. As a consequence, by combining HMMs and ANN, we can expect to take advantage of the robustness and flexibility of the HMMs generative models and of the discriminative power of the ANN. Training such a hybrid system is not straightforward, this is why so few attempts are encountered in literature. We compare several different training schemes: maximum likelihood (ML) and maximum mutual information (MMI) criteria in the framework of online handwriting recognition with a global optimization approach defined at the word level. A new generic criterion mixing generative model and discriminant trainings is proposed, it allows to train a multistate TDNN-HMM system directly at the word level. This architecture is based on an analytical approach with an implicit segmentation. To control the implicit segmentation and to initialize correctly the system without bootstrapping with another recognition system, we have defined a process that constraints the segmentation path and a measure called Average Segmentation Rate (ASR). Recognition experiments on the online IRONOFF database demonstrated the interest of the generic training criterion and the control of the implicit segmentation.

Dissimilarity-Based Classification of Multidimensional Signals by Conjoint Elastic Matching: Application to Phytoplanktonic Species Recognition

The paper describes a classification method of multidimensional signals, based upon a dissimilarity measure between signals. Each new signal is compared to some reference signals through a conjoint dynamic time warping algorithm of their time features series, of which proposed cost function gives out a normalized dissimilarity degree. The classification then consists in presenting these degrees to a classifier, like k-NN, MLP or SVM. This recognition scheme is applied to the automatic estimation of the Phytoplanktonic composition of a marine sample from cytometric curves. At present, biologists are used to a manual classification of signals, that consists in a visual comparison of Phytoplanktonic profiles. The proposed method consequently provides an automatic process, as well as a similar comparison of the signal shapes. We show the relevance of the proposed dissimilarity-based classifier in this environmental application, and compare it with classifiers based on the classical DTW cost-function and also with features-based classifiers.

Comparative study on supervised learning methods for identifying phytoplankton species

Phytoplankton plays an important role in marine ecosystem. It is defined as a biological factor to assess marine quality. The identification of phytoplankton species has a high potential for monitoring environmental, climate changes and for evaluating water quality. However, phytoplankton species identification is not an easy task owing to their variability and ambiguity due to thousands of micro and pico-plankton species. Therefore, the aim of this paper is to build a framework for identifying phytoplankton species and to perform a comparison on different features types and classifiers. We propose a new features type extracted from raw signals of phytoplankton species. We then analyze the performance of various classifiers on the proposed features type as well as two other features types for finding the robust one. Through experiments, it is found that Random Forest using the proposed features gives the best classification results with average accuracy up to 98.24%.

Hybrid Hidden Markov Model for Marine Environment Monitoring

Phytoplankton is an important indicator of water quality assessment. To understand phytoplankton dynamics, many fixed buoys and ferry boxes were implemented, resulting in the generation of substantial data signals. Collected data are used as inputs of an effective monitoring system. The system, based on unsupervised hidden Markov model (HMM), is designed not only to detect phytoplancton blooms but also to understand their dynamics. HMM parameters are usually estimated by an iterative expectation-maximization (EM) approach. We propose to estimate HMM parameters by using spectral clustering algorithm. The monitoring system is assessed based on database signals from MAREL-Carnot station, Boulogne-sur-Mer, France. Experimental results show that the proposed system is efficient to detect environmental states such as phytoplankton productive and nonproductive periods without a priori knowledge. Furthermore, discovered states are consistent with biological interpretation.

A New Fuzzy Logic-Based Similarity Measure Applied to Large Gap Imputation for Uncorrelated Multivariate Time Series

The completion of missing values is a prevalent problem in many domains of pattern recognition and signal processing. Analyzing data with incompleteness may lead to a loss of power and unreliable results, especially for large missing subsequence(s). Therefore, this paper aims to introduce a new approach for filling successive missing values in low/uncorrelated multivariate time series which allows managing a high level of uncertainty. In this way, we propose using a novel fuzzy weighting-based similarity measure. The proposed method involves three main steps. Firstly, for each incomplete signal, the data before a gap and the data after this gap are considered as two separated reference time series with their respective query windows and . We then find the most similar subsequence ( ) to the subsequence before this gap and the most similar one ( ) to the subsequence after the gap . To find these similar windows, we build a new similarity measure based on fuzzy grades of basic similarity measures and on fuzzy logic rules. Finally, we fill in the gap with average values of the window following and the one preceding . The experimental results have demonstrated that the proposed approach outperforms the state-of-the-art methods in case of multivariate time series having low/noncorrelated data but effective information on each signal.

Towards Chl-a bloom understanding by EM-based unsupervised event detection

Marine water quality monitoring and subsequent management require to know when a specific event like harmful algae bloom may occur and which environmental conditions and pressures lead to this event. So, event detection and its dynamic understanding are crucial to adapt strategy. An algorithm is proposed to identify curves mixture and their dynamics features — initiation, duration, peaks and ends of the event. The approach is fully unsupervised, it requires no tuning parameters and is based on Expectation Maximization process to estimate the most robust mixture according to fixed criteria. A complete framework is proposed to deal with a univariate time series with missing data. The approach is applied on Chlorophyll-a series collected weekly since 1989. Chlorophyll-a is a proxy of the phytoplankton biomass. The results are promising according to the phytoplankton composition knowledge, collected at lower frequency, and allowing to discuss about the annual variability of phytoplankton dynamics.

Which DTW method applied to marine univariate time series imputation

Missing data are ubiquitous in any domains of applied sciences. Processing datasets containing missing values can lead to a loss of efficiency and unreliable results, especially for large missing sub-sequence(s). Therefore, the aim of this paper is to build a framework for filling missing values in univariate time series and to perform a comparison of different similarity metrics used for the imputation task. This allows to suggest the most suitable methods for the imputation of marine univariate time series. In the first step, the missing data are completed on various mono-dimensional time series. To fill a missing sub-sequence (gap) in a time series, we first find the most similar sub-sequence to the sub-sequence before (resp. after) this gap according a Dynamic Time Warping (DTW)-cost. Then we complete the gap by the next (resp. previous) sub-sequence of the most similar one. Through experiments results on 5 different datasets we conclude that i) DTW gives the best results when considering the accuracy of imputation values and ii) Adaptive Feature Based DTW (AFBDTW) metric yields very similar shape of imputation values similar to the one of true values.