Romuald Boné

A new boosting algorithm for improved time-series forecasting with recurrent neural networks

Ensemble methods for classification and regression have focused a great deal of attention in recent years. They have shown, both theoretically and empirically, that they are able to perform substantially better than single models in a wide range of tasks. We have adapted an ensemble method to the problem of predicting future values of time series using recurrent neural networks (RNNs) as base learners. The improvement is made by combining a large number of RNNs, each of which is generated by training on a different set of examples. This algorithm is based on the boosting algorithm where difficult points of the time series are concentrated on during the learning process however, unlike the original algorithm, we introduce a new parameter for tuning the boosting influence on available examples. We test our boosting algorithm for RNNs on single-step-ahead and multi-step-ahead prediction problems. The results are then compared to other regression methods, including those of different local approaches. The overall results obtained through our ensemble method are more accurate than those obtained through the standard method, backpropagation through time, on these datasets and perform significantly better even when long-range dependencies play an important role.

Region-Based 2D Deformable Generalized Cylinder for Narrow Structures Segmentation

In this paper, we present a region-based deformable cylinder model, extending the work on classical region-based active contours and gradient-based ribbon snakes. Defined by a central curve playing the role of the medial axis and a variable thickness, the model is endowed with a region-dependent term.This energy follows the narrow band principle, in order to handle local region properties while overcoming limitations of classical edge-based models. The energy is subsequently transformed and derived in order to allow implementation on a polygonal line deformed with gradient descent. The model is used to extract path-like objects in medical and aerial images.

Boosting Recurrent Neural Networks for Time Series Prediction

We adapt a boosting algorithm to the problem of predicting future values of time series, using recurrent neural networks as base learners. The experiments we performed show that boosting actually provides improved results and that the weighted median is better for combining the learners than the weighted mean.

Predicting chaotic time series by boosted recurrent neural networks

This paper discusses the use of a recent boosting algorithm for recurrent neural networks as a tool to model nonlinear dynamical systems. It combines a large number of RNNs, each of which is generated by training on a different set of examples. This algorithm is based on the boosting algorithm where difficult examples are concentrated on during the learning process. However, unlike the original algorithm, all examples available are taken into account. The ability of the method to internally encode useful information on the underlying process is illustrated by several experiments on well known chaotic processes. Our model is able to find an appropriate internal representation of the underlying process from the observation of a subset of the states variables. We obtain improved prediction performances.

Greedy Algorithm and Physics-Based Method for Active Contours and Surfaces: A Comparative Study

Deformable models, such as the discrete active contour and surface, imply the use of iterative evolution methods to perform 2D and 3D image segmentation. Among the several existing evolution methods, we focus on the greedy algorithm, which minimizes an energy functional, and the physics-based method, which applies forces in order to solve a dynamic differential equation. In this paper, we compare the greedy and physics-based approaches applied on 2D and 3D models, as regards overall speed and segmentation quality, quantified with an evaluating function mainly based on the mean distance between the model and the desired shape.

Study of the behavior of a new boosting algorithm for recurrent neural networks

We present an algorithm for improving the accuracy of recurrent neural networks (RNNs) for time series forecasting. The improvement is achieved by combining a large number of RNNs, each of them is generated by training on a different set of examples. This algorithm is based on the boosting algorithm and allows concentrating the training on difficult examples but, unlike the original algorithm, by taking into account all the available examples. We study the behavior of our method applied on three time series of reference with three loss functions and with different values of a parameter. We compare the performances obtained with other regression methods.

Active Contours Driven by Supervised Binary Classifiers for Texture Segmentation

In this paper, we propose a new active contour model for supervised texture segmentation driven by a binary classifier instead of a standard motion equation. A recent level set implementation developed by Shi et al in [1] is employed in an original way to introduce the classifier in the active contour. Carried out on a learning image, an expert segmentation is used to build the learning dataset composed of samples defined by their Haralick texture features. Then, the pre-learned classifier is used to drive the active contour among several test images. Results of three active contours driven by binary classifiers are presented: a k-nearest-neighbors model, a support vector machine model and a neural network model. Results are presented on medical echographic images and remote sensing images and compared to the Chan-Vese region-based active contour in terms of accuracy, bringing out the high performances of the proposed models.

Advanced Methods for Time Series Prediction Using Recurrent Neural Networks

Time series prediction has important applications in various domains such as medicine, ecology, meteorology, industrial control or finance. Generally the characteristics of the phenomenon which generates the series are unknown. The information available for the prediction is limited to the past values of the series. The relations which describe the evolution should be deduced from these values, in the form of functional relation approximations between the past and the future values. The most usually adopted approach to consider the future values ( ) 1 t x + consists in using a function f which takes as input a time window of fixed size M representing the recent history of the time series. ( ) ( ) ( ) ( ) ( ) [ ] τ − − τ − = 1 M t x , , t x , t x t ... x (1) ( ) ( ) ( ) t f t x x = τ + (2)

Hierarchical Mesh Segmentation Using Waterfall and Dynamics

Mesh segmentation is a fundamental problem in computer graphics and has become an important component in many applications. This paper proposes a new hierarchical mesh segmentation method based on waterfall and dynamics. Watershed transformation allows segmenting the mesh in small patches and dynamics provide information about boundaries and merging possibilities. From region and boundary information, a hierarchical process based on waterfall is computed in order to build the merging tree. This tree contains all the schemes of segmentation and the user can easily browse the different levels of segmentation.

A supervised texture-based active contour model with linear programming

In this paper we propose a new supervised active contour model evolving with Haralick texture features. This model is divided in two stages. First, we use a supervised step where the user defines an ideal segmentation on a learning image. A linear programming model, modeling the behavior of the active contour, is then used to determine the weights of the Haralick features leading to the optimal segmentation. In a second step, a texture-oriented active contour based on the Chan-Vese model is launched on several test images with the learned weights and the closest segmentations to the one defined on the learning image is determined. Results of our method are presented on medical echographic images.