Achille J.-P. Braquelaire

Segmentation-based multi-class semantic object detection

In this paper we study the problem of the detection of semantic objects from known categories in images. Unlike existing techniques which operate at the pixel or at a patch level for recognition, we propose to rely on the categorization of image segments. Recent work has highlighted that image segments provide a sound support for visual object class recognition. In this work, we use image segments as primitives to extract robust features and train detection models for a predefined set of categories. Several segmentation algorithms are benchmarked and their performances for segment recognition are compared. We then propose two methods for enhancing the segments classification, one based on the fusion of the classification results obtained with the different segmentations, the other one based on the optimization of the global labelling by correcting local ambiguities between neighbor segments. We use as a benchmark the Microsoft MSRC-21 image database and show that our method competes with the current state-of-the-art.

Intuitive shape modeling by shading design

Shading has a great impact to the human perception of 3D objects. Thus, in order to create or to deform a 3D object, it seems natural to manipulate its perceived shading. This paper presents a new solution for the software implementation of this idea. Our approach is based on the ability of a user to coarsely draw a shading, under different lighting directions. With this intuitive process, users can create or edit a height field (locally or globally), that will correspond to the drawn shading values. Moreover, we present the possibility to edit the shading intensity by means of a specular reflectance model.

Comparison and convergence of two topological models for 3D image segmentation

In this paper we compare two topological models of 3D segmented images representation. These models are based on a collaboration between a topological representation and a geometrical representation of the regions of the segmented image. In both models the description of the topology lays on topological maps with a combinatorial representation. A geometrical embedding of maps is used to describe the geometry of regions. Both models differ in the way of defining this embedding. The aim of this paper is to compare these two models from the point of view of their interest for image segmentation, and to explore the possibility and the interest of making both these models converge.

Reconstruction of Lambertian surfaces by discrete equal height contours and regions propagation

This paper describes two new methods for the reconstruction of discrete surfaces from shading images. Both approaches are based on the reconstruction of a discrete surface by mixing photometric and geometric techniques. The processing of photometric information is based on reflectance maps, which are classic tools of Shape from Shading. The geometric features are extracted from the discrete surface and propagated along the surface. The propagation is based in one case on equal height discrete contour propagation and in the other case on region propagation. Both methods allow photometric stereo. Results of reconstruction from synthetic and real images are presented.

3D image topological structuring with an oriented boundary graph for split and merge segmentation

In this paper, we present a new representation model for the topology and the geometry of a 3D segmented image. This model has been designed to provide main features and operations required by a 3D image segmentation library. It is mainly devoted to region based segmentation methods such as split and merge algorithms but is also convenient for contour based approaches. The model has been fully implemented and tested both on synthetic and real 3D images.

Representing and segmenting 2d images by means of planar maps with discrete embeddings: from model to applications

Representing the regions of a segmented image is an important aspect of image segmentation. Several different models have been proposed to represent the regions of a segmented image but most of them are dedicated to a specific method. Among the non hierarchical models, the model of planar maps with discrete embedding is certainly the most versatile one. Maps have the great advantage to provide a continuity of representation from the abstract mathematical model to the concrete implementation. They encode and provide most of topological and geometrical features required by segmentation algorithms and can be efficiently updated. In this paper we give an overview of the use of planar maps with discrete embedding in the context of image segmentation and we show how to design, implement and use a general environment for 2D image segmentation, from the mathematical model up to a real application. The model, data structure, algorithms and API described in this paper are currently implemented in a software which will be available under LGPL in the course of year 2005.

Reconstruction of Discrete Surfaces from Shading Images by Propagation of Geometric Features

This paper describes two new methods for the reconstruction of discrete surfaces from shading images. Both approaches are based on the reconstruction of a discrete surface by mixing photometric and geometric techniques. The processing of photometric informations is based on reflectance maps which are classic tools of shape from shading. The geometric features are extracted from the discrete surface and propagated along the surface. The propagation is based in one case on equal height discrete contour propagation and in the other case on region propagation. Both methods allow photometric stereo. Results of reconstruction from synthetic and real images are presented.

ESPI Image Indexing and Similarity Search in Radon Transform Domain

Content Based Image Retrieval is a topic which has received a lot of attention and increasing popularity due to a wide range of applications. In this paper, we present a similarity measure for CBIR in an industrial context, where the images of a vibration phenomenon are obtained by Electronic Speckle Pattern Interferometry. Images obtained have very poor visual characteristics, and traditional CBIR systems which rely on color or texture information could not be efficient. We propose a CBIR approach based on the 1-dimensional projections of the images obtained by the Radon transform. Experiments show that this signature is relevant and enables good retrieval performances compared to a baseline image correlation.

Oriented Boundary Graph: A Framework to Design and Implement 3D Segmentation Algorithms

In this paper we show the interest of a topological model to represent 3D segmented image which is a good compromise between the complete but time consuming representations and the partial but not expressive enough ones. We show that this model, called Oriented Boundary Graph, provides an effective framework for both volumic image analysis and segmentation. The Oriented Boundary Graph provides an efficient implementation of a set of primitives suitable for the design complex segmentation algorithms and to implement the computation of the segmented image characteristics needed by such algorithms. We first present the framework and give the time complexity of its main primitives. Then, we give some examples of the use of this framework in order to efficiently design non-trivial image analysis operations and image segmentation algorithms. Those examples are applied on 3D CT-scan data.

A reversible and statistical method for discrete surfaces smoothing

In this article we propose an original reversible method for discrete surface smoothing. This method is based on a statistical estimation of the discrete tangent plane on the voxels of the discrete surface. A geometrical constraint is used to control the recognition of the tangent plane. The resulting surface representation allows us to get both smooth normal vectors of the surface and a smooth surface mesh while preserving the geometrical properties of the surface.