Muriel Visani

NAVIGALA: AN ORIGINAL SYMBOL CLASSIFIER BASED ON NAVIGATION THROUGH A GALOIS LATTICE

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.

The ICDAR/GREC 2013 Music Scores Competition: Staff Removal

The first competition on music scores that was organized at ICDAR and GREC in 2011 awoke the interest of researchers, who participated in both staff removal and writer identification tasks. In this second edition, we focus on the staff removal task and simulate a real case scenario concerning old and degraded music scores. For this purpose, we have generated a new set of semi-synthetic images using two degradation models that we previously introduced: local noise and 3D distortions. In this extended paper we provide an extended description of the dataset, degradation models, evaluation metrics, the participant’s methods and the obtained results that could not be presented at ICDAR and GREC proceedings due to page limitations.

An experimental comparison of clustering methods for content-based indexing of large image databases

In recent years, the expansion of acquisition devices such as digital cameras, the development of storage and transmission techniques of multimedia documents and the development of tablet computers facilitate the development of many large image databases as well as the interactions with the users. This increases the need for efficient and robust methods for finding information in these huge masses of data, including feature extraction methods and feature space structuring methods. The feature extraction methods aim to extract, for each image, one or more visual signatures representing the content of this image. The feature space structuring methods organize indexed images in order to facilitate, accelerate and improve the results of further retrieval. Clustering is one kind of feature space structuring methods. There are different types of clustering such as hierarchical clustering, density-based clustering, grid-based clustering, etc. In an interactive context where the user may modify the automatic clustering results, incrementality and hierarchical structuring are properties growing in interest for the clustering algorithms. In this article, we propose an experimental comparison of different clustering methods for structuring large image databases, using a rigorous experimental protocol. We use different image databases of increasing sizes (Wang, PascalVoc2006, Caltech101, Corel30k) to study the scalability of the different approaches.

Unsupervised word spotting using a graph representation based on invariants

We are currently working on the concept of an interactive word retrieval system for ancient document collection navigation, based on query composition for non-expert users. We have introduced a new notion: invariants, which are writing pieces automatically extracted from the old document collection. The invariants can be used in query making process in where the user selects and composes appropriate invariants to make the query. The invariants can be also used as descriptor to characterize word images. We introduced our unsupervised method for extracting invariants in our earlier paper. In this paper, we present a new structural word spotting system using a graph representation based on invariants as a descriptor. Through experiments, we conclude that our proposed system can adapt to different types of homogenous alphabetic languages documents (regardless of language/script, antiquity, handwritten or printed).

A new interactive semi-supervised clustering model for large image database indexing

Indexing methods play a very important role in finding information in large image databases. They organize indexed images in order to facilitate, accelerate and improve the results for later retrieval. Alternatively, clustering may be used for structuring the feature space so as to organize the dataset into groups of similar objects without prior knowledge (unsupervised clustering) or with a limited amount of prior knowledge (semi-supervised clustering). In this paper, we introduce a new interactive semi-supervised clustering model where prior information is integrated via pairwise constraints between images. The proposed method allows users to provide feedback in order to improve the clustering results according to their wishes. Different strategies for deducing pairwise constraints from user feedback were investigated. Our experiments on different image databases (Wang, PascalVoc2006, Caltech101) show that the proposed method outperforms semi-supervised HMRF-kmeans (Basu et al., 2004).

Writer Identification Using TF-IDF for Cursive Handwritten Word Recognition

In this paper, we present two text-independent writer identification methods in a closed-world context. Both methods use on-line and off-line features jointly with a classifier inspired from information retrieval methods. These methods are local, respectively based on the character and grapheme levels. This writer identification engine may be used to personalize our cursive word recognition engine~\cite{icfhr2010} to the handwriting style of the writer, resulting in an adaptive cursive word recognizer. Experiments assess the effectiveness of the proposed approaches in a context of writer identification as well as integrated to our cursive word recognizer to make it adaptive.

Document Retrieval Based on Logo Spotting Using Key-Point Matching

In this paper, we present an approach to retrieve documents based on logo spotting and recognition. A document retrieval system is proposed inspired from our previous method for logo spotting and recognition. First, the key-points from both the query logo images and a given set of document images are extracted and described by SIFT descriptor, and are matched in the SIFT feature space. They are filtered by the nearest neighbor matching rule based on the two nearest neighbors and are then post-filtered with BRIEF descriptor. Secondly, logo segmentation is performed using spatial density-based clustering, and homography is used to filter the matched key-points as a post processing. Finally, for ranking, we use two measures which are calculated based on the number of matched key-points. Tested on a well-known benchmark database of real world documents containing logos Tobacco-800, our approach achieves better performance than the state-of-the-art methods.

On the Joint Use of a Structural Signature and a Galois Lattice Classifier for Symbol Recognition

In this paper, we propose a new approach for symbol recognition using structural signatures and a Galois Lattice as classifier. The structural signatures are based on topological graphs computed from segments which are extracted from the symbol images by using an adapted Hough transform. These structural signatures, which can be seen as dynamic paths which carry high level information, are robust towards various transformations. They are classified by using a Galois Lattice as a classifier. The performances of the proposed approach are evaluated on the GREC03 symbol database and the experimental results we obtain are encouraging.

Semi-synthetic Document Image Generation Using Texture Mapping on Scanned 3D Document Shapes

This article presents a method for generating semi-synthetic images of old documents where the pages might be torn (not flat). By using only 2D deformation models, most existing methods give non-realistic synthetic document images. Thus, we propose to use 3D approach for reproducing geometric distortions in real documents. First, a new proposed texture coordinate generation technique extracts texture coordinates of each vertex in the document shape (mesh) resulting from 3D scanning of a real degraded document. Then, any 2D document image can be overlayed on the mesh by using an existing texture image mapping method. As a result, many complex real geometric distortions can be integrated in generated synthetic images. These images then can be used for enriching training sets or for performance evaluation. The degradation method here is jointly used with the character degradation model we proposed in [1] to generate the 6000 semi-synthetic degraded images of the music score removal staff line competition of ICDAR 2013.

Text and non-text segmentation based on connected component features

Document image segmentation is crucial to OCR and other digitization processes. In this paper, we present a learning-based approach for text and non-text separation in document images. The training features are extracted at the level of connected components, a mid-level between the slow noise-sensitive pixel level, and the segmentation-dependent zone level. Given all types, shapes and sizes of connected components, we extract a powerful set of features based on size, shape, stroke width and position of each connected component. Adaboosting with Decision trees is used for labeling connected components. Finally, the classification of connected components into text and non-text is corrected based on classification probabilities and size as well as stroke width analysis of the nearest neighbors of a connected component. The performance of our approach has been evaluated on the two standard datasets: UW-III and ICDAR-2009 competition for document layout analysis. Our results demonstrate that the proposed approach achieves competitive performance for segmenting text and non-text in document images of variable content and degradation.