Thierry Géraud

The SCRIBO module of the Olena platform: A free software framework for document image analysis

Abstract

Electronic documents are being more and more usable thanks to better and more affordable network, storage and computational facilities. But in order to benefit from computer-aided document management, paper documents must be digitized and analyzed. This task may be challenging at several levels. Data may be of multiple types thus requiring different adapted processing chains. The tools to be developed should also take into account the needs and knowledge of users, ranging from a simple graphical application to a complete programming framework. Finally, the data sets to process may be large. In this paper, we expose a set of features that a Document Image Analysis framework should provide to handle the previous issues. In particular, a good strategy to address both flexibility and efficiency issues is the Generic Programming (GP) paradigm. These ideas are implemented as an open source module, SCRIBO, built on top of Olena, a generic and efficient image processing platform. Our solution features services such as preprocessing filters, text detection, page segmentation and document reconstruction (as XML, PDF or HTML documents). This framework, composed of reusable software components, can be used to create full-fledged graphical applications, small utilities, or processing chains to be integrated into third-party projects.

Continue reading

Une approche générique du logiciel pour le traitement d’images préservant les performances

Abstract

De plus en plus d’outils logiciels modernes pour le traitement d’images sont conçus en prenant en compte le problème de la généricité du code, c’est-à-dire la possibilité d’écrire des algorithmes réutilisables, compatibles avec de nombreux types d’entrées. Cependant, ce choix de conception se fait souvent au détriment des performances du code exécuté. Du fait de la grande variété des types d’images existants et de la nécessité d’avoir des implémentations rapides, généricité et performance apparaissent comme des qualités essentielles du logiciel en traitement d’images. Cet article présente une approche préservant les performances dans un framework logiciel générique tirant parti des caractéristiques des types de données utilisés. Grâce à celles-ci, il est possible d’écrire des variantes d’algorithmes génériques offrant un compromis entre généricité et performance. Ces alternatives sont capables de préserver une partie des aspects génériques d’origine tout en apportant des gains substantiels à l’exécution. D’après nos essais, ces optimisations génériques fournissent des performances supportant la comparaison avec du code dédié, allant parfois même jusqu’à surpasser des routines optimisées manuellement.

Continue reading

Morphologie et algorithmes

Abstract

Continue reading

Algorithms for mathematical morphology

Abstract

Continue reading

Why and how to design a generic and efficient image processing framework: The case of the Milena library

Abstract

Most image processing frameworks are not generic enough to provide true reusability of data structures and algorithms. In fact, genericity allows users to write and experiment virtually any method on any compatible input(s). In this paper, we advocate the use of generic programming in the design of image processing software, while preserving performances close to dedicated code. The implementation of our proposal, Milena, a generic and efficient library, illustrates the benefits of our approach.

Continue reading

Milena: Write generic morphological algorithms once, run on many kinds of images

Abstract

We present a programming framework for discrete mathematical morphology centered on the concept of genericity. We show that formal definitions of morphological algorithms can be translated into actual code, usable on virtually any kind of compatible images, provided a general definition of the concept of image is given. This work is implemented in Milena, a generic, efficient, and user-friendly image processing library.

Continue reading

Semantics-driven genericity: A sequel to the static C++ object-oriented programming paradigm (SCOOP 2)

Abstract

Classical (unbounded) genericity in 03 defines the interactions between generic data types and algorithms in terms of concepts. Concepts define the requirements over a type (or a parameter) by expressing constraints on its methods and dependent types (typedefs). The upcoming 0x standard will promote concepts from abstract entities (not directly enforced by the tools) to language constructs, enabling compilers and tools to perform additional checks on generic constructs as well as enabling new features (e.g., concept-based overloading). Most modern languages support this notion of signature on generic types. However, generic types built on other types and relying on concepts to both ensure type conformance and drive code specialization, restrain the interface and the implementation of the newly created type: specific methods and associated types not mentioned in the concept will not be part of the new type. The paradigm of concept-based genericity lacks the required semantics to transform types while retaining or adapting their intrinsic capabilities. We present a new form of semantically-enriched genericity allowing static generic type transformations through a simple form of type introspection based on type metadata called properties. This approach relies on a new Static Object-Oriented Programming (SCOOP) paradigm, and is adapted to the creation of generic and efficient libraries, especially in the field of scientific computing. Our proposal uses a metaprogramming facility built into a library called Static, and doesn’t require any language extension nor additional processing (preprocessor, transformation tool).

Continue reading

Algorithme de calcul de l’arbre des composantes avec applications à la reconnaissance des formes en imagerie satellitaire

Abstract

In this paper a new algorithm to compute the component tree is presented. As compared to the state-of-the-art, this algorithm does not use excessive memory and is able to work efficiently on images whose values are highly quantized or even with images having floating values. We also describe how it can be applied to astronomical data to identify relevant objects.

Continue reading

Effective component tree computation with application to pattern recognition in astronomical imaging

Abstract

In this paper a new algorithm to compute the component tree is presented. As compared to the state of the art, this algorithm does not use excessive memory and is able to work efficiently on images whose values are highly quantized or even with images having floating values. We also describe how it can be applied to astronomical data to identify relevant objects.

Continue reading

Project EFIGI: Automatic classification of galaxies

Abstract

We propose an automatic system to classify images of galaxies with varying resolution. Morphologically typing galaxies is a difficult task in particular for distant galaxies convolved by a point-spread function and suffering from a poor signal-to-noise ratio. In the context of the first phase of the project EFIGI (extraction of the idealized shapes of galaxies in imagery), we present the three steps of our software: cleaning, dimensionality reduction and supervised learning. We present preliminary results derived from a subset of 774 galaxies from the Principal Galaxies Catalog and compare them to human classifications made by astronomers. We use g-band images from the Sloan Digital Sky Survey. Finally, we discuss future improvements which we intend to implement before releasing our tool to the community.

Continue reading