Florence Cloppet

Segmentation of overlapping/aggregating nuclei cells in biological images

This paper presents a method of overlapping/aggregating nuclei cells segmentation. This method is based on the watershed segmentation algorithm, but the specificity of this work is to introduce some prior information about the usual shape of normal/pathological nuclei cells. Such prior information will help to optimize the right set of markers, from which the flooding will be done. This approach has been implemented and tested, and the results evaluated by cell biology experts show the efficiency of the proposed approach.

Segmentation of complex nucleus configurations in biological images

This paper presents a new segmentation method of complex nucleus configurations. The specificity of this work is to introduce prior information about the usual shape of cells nuclei, in order to optimize the selection of markers from which the flooding will start, during the watershed-based segmentation.

Feature selection combining genetic algorithm and Adaboost classifiers

This paper presents a fast method using simple genetic algorithms (GAs) for features selection. Unlike traditional approaches using GAs, we have used the combination of Adaboost classifiers to evaluate an individual of the population. So, the fitness function we have used is defined by the error rate of this combination. This approach has been implemented and tested on the MNIST database and the results confirm the effectiveness and the robustness of the proposed approach.

Angular bisector network, a simplified generalized Voronoi diagram: application to processing complex intersections in biomedical images

One of the major goals of computer vision is the research and the development of flexible methods for shape description. A large group of shape description techniques is given by heuristic approaches, which yield acceptable results in the description of simple shapes and regions. In this case, objects are represented by a planar graph with nodes symbolizing subregions from region decomposition, and region shape is then described by the graph properties. In the paper, the angular bisector network (ABN), a descriptor of polygonal shape, is used to automatically detect intersections between neurites of cell structures. Some properties of the ABN, such as linear algebraic complexity, easy extraction of characteristic points, etc., are very useful and experimental results are promising.

NBC: A Workstation for Biological Neural Network Simulation

Neuro_bio_clusters (NBC) is a software package created to simulate interacting biological neural networks. It is designed for neuroscientists who know little about computer sciences. NBC provides two neuron simulation levels: a simple, fast phenomenological one in which membrane potential and threshold are included, and a more sophisticated, slower one, based on conductance variations in different ionic channels. NBC enables the simulation of biologically plausible networks formed by several interconnected neural clusters connected through pathways of variable length, which can receive external inputs and whose connection matrix is specified. It also enables the input, the anatomical characteristics, and other properties of individual units and synapses to be modified during the simulation thereby taking into account the changes in environmental conditions due to learning or pathological conditions. NBC provides an analysis tool highly useful in the study of network behavior, at both global and unitary levels, and in both the frequency and the temporal domain. NBC is menu driven with a user-friendly XWindow/Motif interface and produces graphic outputs.

First Combined in Vivo X-Ray Tomography and High-Resolution Molecular Electron Paramagnetic Resonance (EPR) Imaging of the Mouse Knee Joint Taking into Account the Disappearance Kinetics of the EPR Probe:

Although laboratory data clearly suggest a role for oxidants (dioxygen and free radicals derived from dioxygen) in the pathogenesis of many age-related and degenerative diseases (such as arthrosis and arthritis), methods to image such species in vivo are still very limited. This methodological problem limits physiopathologic studies about the role of those species in vivo, the effects of their regulation using various drugs, and the evaluation of their levels for diagnosis of degenerative diseases. In vivo electron paramagnetic resonance (EPR) imaging and spectroscopy are unique, noninvasive methods used to specifically detect and quantify paramagnetic species. However, two problems limit their application: the anatomic location of the EPR image in the animal body and the relative instability of the EPR probes. Our aim is to use EPR imaging to obtain physiologic and pathologic information on the mouse knee joint. This article reports the first in vivo EPR image of a small tissue, the mouse knee joint, with good resolution (≈ 160 μm) after intra-articular injection of a triarylmethyl radical EPR probe. It was obtained by combining EPR and x-ray micro-computed tomography for the first time and by taking into account the disappearance kinetics of the EPR probe during image acquisition to reconstruct the image. This multidisciplinary approach opens the way to high-resolution EPR imaging and local metabolism studies of radical species in vivo in different physiologic and pathologic situations.

Graphical drop caps indexing

This paper presents a method for graphical drop caps indexing. Drop caps are extracted from old books. Finding a method classifying them according to styles defined by the historian is of considerable interest. The developed method is a statistical approach, where all possible patterns included in a pixel mask are processed in order to extract indexes that characterize the image. Then these indexes are used to classify a query drop cap by searching its most similar drop caps in the indexed base.

Detection of linear structures in biological images

We present a method for detecting linear structures in biological microscopy images. The contribution of this paper is to unify the Beamlet transform with linear filtering techniques and propose a new detector, the feature-adpated beamlet transform. Our detector is able to incorporate knowledge about the desired line-profile lying along curves, like edges or ridges. We propose an efficient implementation in the case the profile is designed as a steerable filter. Preliminary results on biological images containing both edge- and ridge-like profiles have shown significant improvements over linear detector techniques and multiscale detection techniques based on traditional Beamlet transform.

Visual Perception Driven Registration of Mammograms

This paper aims to develop a methodology to register pairs of temporal mammograms. Control points based on anatomical features are detected in an automated way. Thereby, image semantic is used to extract landmarks based on these control points. A referential is generated from these control points based on this referential the studied images are realigned using different levels of observation leading to both rigid and pseudo non-rigid transforms according to expert mammogram reading.

Change detection in aerial images

This paper deals with how to characterize texture and how to get a good description of images with a minimal number of parameters. This procedure is more objective than textual data. Texture characterization has been used in a matching system to detect changes in couples of aerial images taken at two different times using different order of statistics to describe images. The results are quite encouraging.