We present eBDtheque, a database of various comic book images and their ground truth for panels, balloons and text lines plus semantic annotations. The database consists of a hundred pages of various comic book albums, Franco-Belgian, American comics and mangas. Additionally, we present the piece of software used to establish the ground truth and a tool to validate results against this ground truth. Everything is publicly available for scientific use on http://ebdtheque.univ-lr.fr.

International audienceThe notion of dependencies between "attributes" arises in many areas such as relational databases, data analysis, data-mining, formal concept analysis, knowledge spaces. Formalization of dependencies leads to the notion of so-called full implicational systems (or full family of functional dependencies) which is in one-to-one correspondence with the other signicant notions of closure operator and of closure system. An efficient generation of a full implicational system (or of a closure system) can be performed from equivalent implicational systems and in particular from bases for such systems, for example, the so-called canonical basis. This paper shows the equality between five other bases originating from dfferent works and satisfying various properties (in particular they are unit implicational systems). The three main properties of this unique basis are the directness, canonical and minimal properties, whence the name canonical direct unit implicational basis given to this unit implicational system. The paper also gives a characterization of this canonical basis and it makes precise its link with the prime implicants of the Horn function associated to a closure operator.La notion de dépendance entre attributs est présente dans de nombreux domaines comme les bases de données relationnelles, l'analyse et la fouille des données, l'analyse formelle des concepts ou les espaces de connaissance. La formalisation de ces dépendances conduit à la notion de système complet d'implications, notion en correspondance biunivoque avec les notions d'opérateur de fermeture ou de famille de Moore. Un système complet d'implications (ou une fermeture) peut être engendré efficacement à partir de bases, par exemple la la base canonique (de Duquenne et Guigues). Dans ce texte, nous montrons l'identité de cinq autres bases obtenues dans des domaines variés . Cette base est unitaire directe, canonique et minimale, d'où son nom de base unitaire canonique directe. Nous donnons une caractérisation de cette base et nous montrons sa correspondance avec les implicants premiers de la fonction de Horn associée à une fermeture

No abstract

This paper deals with a supervised classification method, using Galois Lattices based on a navigation-based strategy. Coming from the field of data mining techniques, most literature on the subject using Galois lattices relies on selection-based strategies, which consists of selecting/choosing the concepts which encode the most relevant information from the huge amount of available data. Generally, the classification step is then processed by a classical classifier such as the k-nearest neighbors rule or the Bayesian classifier. Opposed to these selection-based strategies are navigation-based approaches which perform the classification stage by navigating through the complete lattice (similar to the navigation in a classification tree), without applying any selection operation. Our approach, named Navigala, proposes an original navigation-based approach for supervised classification, applied in the context of noisy symbol recognition. Based on a state of the art dealing with Galois Lattices classification based methods, including a comparison between possible selection and navigation strategies, this paper proposes a description of NAVIGALA and its implementation in the context of symbol recognition. Some objective quantitative and qualitative evaluations of the approach are proposed, in order to highlight the relevance of the method.

Since the beginning of the twenty-first century, the cultural industry has been through a massive and historical mutation induced by the rise of digital technologies. The comic books industry keeps looking for the right solution and has not yet produced anything as convincing as the music or movie have. A lot of energy has been spent to transfer printed material to digital supports so far. The specificities of those supports are not always exploited at the best of their capabilities, while they could potentially be used to create new reading conventions. In spite of the needs induced by the large amount of data created since the beginning of the comics history, content indexing has been left behind. It is indeed quite a challenge to index such a composition of textual and visual information. While a growing number of researchers are working on comic books' image analysis from a low-level point of view, only a few are tackling the issue of representing the content at a high semantic level. We propose in this article a framework to handle the content of a comic book, to support the automatic extraction of its visual components and to formalize the semantic of the domain's codes. We tested our framework over two applications: 1) the unsupervised content discovery of comic books' images, 2) its capabilities to handle complex layouts and to produce a respectful browsing experience to the digital comics reader.