Gemma Sánchez - Academia.edu (original) (raw)

Papers by Gemma Sánchez

This article presents a pen-based framework for manual edition of digital documents on tablet com... more This article presents a pen-based framework for manual edition of digital documents on tablet computers. In this system, the user draws certain proofreading symbols on the text parts to edit; some symbols can be accompanied by handwritten text. The input is interpreted and the corresponding editing action is executed in real time. The possibility that the input contains handwritten text is a novelty with respect to previous real-time systems that faced sketch-based edition, where usually text input is carried out via keyboard. Also, multimodal feedback mechanisms for error recovery are present. In this work we focus on the symbol recognition part. Different features are evaluated in recognition experiments using a support vector machine classifier. Experiments show that the symbol recognition is efficient enough for a real-time task and that the system can be used in real conditions with some experience.

Lung, 2006

The effects of training on dynamic hyperinflation in stable chronic obstructive pulmonary disease... more The effects of training on dynamic hyperinflation in stable chronic obstructive pulmonary disease (COPD) were investigated by using a controlled study of 28 subjects with FEV1 = 42.5 (8.3 SD)%pred and 20 matched controls [FEV1 = 44.9 (10.4)%pred]. Training consisted of spending 45 min/day, 4 days/week on a cycle-ergometer for six weeks. Maximal inspiratory and expiratory pressures (MIP and MEP), lung volumes, and two constant-work-rate (CWR) exercise tests (low- and high-intensity) were performed. Significant (p 2O], MEP [+18 (20) cmH2O], and endurance to high-intensity CWR [+7(5) min], and there were significant decreases in respiratory rate and end-expiratory lung volume (EELV) during both exercise tests. At 5 min, EELV decreased 0.1(0.08) L and 0.31(0.13) L and at end of exercise, EELV decreased by 0.09(0.07) L and 0.15(0.11) L respectively, for the moderate- and high-intensity tests. Dyspnea also decreased significantly at both exercise intensities. No changes were observed in the control group. Increased endurance showed independent significant (p

This paper presents an algorithm for recognizing symbols with textured elements in a graphical do... more This paper presents an algorithm for recognizing symbols with textured elements in a graphical document. A region adjacency graph represents the document. The texture symbols are modeled by a graph grammar. An inference algorithm is applied to learn such grammar from an instance of the texture. For recognition, a parsing process is applied. Since documents present distortions, error-correcting rules are added to the grammar.

This paper presents a syntactic recognition approach for on-line drawn graphical symbols. The pro... more This paper presents a syntactic recognition approach for on-line drawn graphical symbols. The proposed method consists in an incremental on-line predictive parser based on symbol descriptions by an adjacency grammar. The parser analyzes input strokes as they are drawn by the user and is able to get ahead which symbols are likely to be recognized when a partial subshape is drawn in an intermediate state. In addition, the parser takes into account two issues. First, symbol strokes are drawn in any order by the user and second, since it is an on-line framework, the system requires real-time response. The method has been applied to an on-line sketching interface for architectural symbols.

This article proposes a novel similarity measure between vector sequences. Recently, a model-base... more This article proposes a novel similarity measure between vector sequences. Recently, a model-based approach was introduced to address this issue. It consists in modeling each sequence with a continuous Hidden Markov Model (C-HMM) and computing a probabilistic measure of similarity between C-HMMs. In this paper we propose to model sequences with semi-continuous HMMs (SC-HMMs): the Gaussians of the SC-HMMs are constrained to belong to a shared pool of Gaussians. This constraint provides two major benefits. First, the a priori information contained in the common set of Gaussians leads to a more accurate estimate of the HMM parameters. Second, the computation of a probabilistic similarity between two SC-HMMs can be simplified to a Dynamic Time Warping (DTW) between their mixture weight vectors, which reduces significantly the computational cost. Experimental results on a handwritten word retrieval task show that the proposed similarity outperforms the traditional DTW between the original sequences, and the model-based approach which uses C-HMMs. We also show that this increase in accuracy can be traded against a significant reduction of the computational cost (up to 100 times). * J.A. Rodríguez-Serrano was a visitor at XRCE and a Ph.D. candidate at the CVC while this work was conducted.

Pattern Analysis and Applications, 2010

In this paper, we address the problem of symbol spotting in technical document images applied to ... more In this paper, we address the problem of symbol spotting in technical document images applied to scanned and vectorized line drawings. Like any information spotting architecture, our approach has two components. First, symbols are decomposed in primitives which are compactly represented and second a primitive indexing structure aims to efficiently retrieve similar primitives. Primitives are encoded in terms of attributed strings representing closed regions. Similar strings are clustered in a lookup table so that the set median strings act as indexing keys. A voting scheme formulates hypothesis in certain locations of the line drawing image where there is a high presence of regions similar to the queried ones, and therefore, a high probability to find the queried graphical symbol. The proposed approach is illustrated in a framework consisting in spotting furniture symbols in architectural drawings. It has been proved to work even in the presence of noise and distortion introduced by the scanning and raster-to-vector processes.

Pattern Recognition Letters, 2009

Many symbol recognition problems require the use of robust descriptors in order to obtain rich in... more Many symbol recognition problems require the use of robust descriptors in order to obtain rich information of the data. However, the research of a good descriptor is still an open issue due to the high variability of symbols appearance. Rotation, partial occlusions, elastic deformations, intra-class and inter-class variations, or high variability among symbols due to different writing styles, are just a few problems. In this paper, we introduce a symbol shape description to deal with the changes in appearance that these types of symbols suffer. The shape of the symbol is aligned based on principal components to make the recognition invariant to rotation and reflection. Then, we present the Blurred Shape Model descriptor (BSM), where new features encode the probability of appearance of each pixel that outlines the symbols shape. Moreover, we include the new descriptor in a system to deal with multi-class symbol categorization problems. Adaboost is used to train the binary classifiers, learning the BSM features that better split symbol classes. Then, the binary problems are embedded in an Error-Correcting Output Codes framework (ECOC) to deal with the multi-class case. The methodology is evaluated on different synthetic and real data sets. State-of-the-art descriptors and classifiers are compared, showing the robustness and better performance of the present scheme to classify symbols with high variability of appearance.

The recognition of symbols in graphic documents is an intensive research activity in the communit... more The recognition of symbols in graphic documents is an intensive research activity in the community of pattern recognition and document analysis. A key issue in the interpretation of maps, engineering drawings, diagrams, etc. is the recognition of domain dependent symbols according to a symbol database. In this work we first review the most outstanding symbol recognition methods from two different points of view: application domains and pattern recognition methods. In the second part of the paper, open and unaddressed problems involved in symbol recognition are described, analyzing their current state of art and discussing future research challenges. Thus, issues such as symbol representation, matching, segmentation, learning, scalability of recognition methods and performance evaluation are addressed in this work. Finally, we discuss the perspectives of symbol recognition concerning to new paradigms such as user interfaces in handheld computers or document database and WWW indexing by graphical content.

In this paper we present an innovative approach to automatically generate adjacency grammars desc... more In this paper we present an innovative approach to automatically generate adjacency grammars describing graphical symbols. A grammar production is formulated in terms of rulesets of geometrical constraints among symbol primitives. Given a set of symbol instances sketched by a user using a digital pen, our approach infers the grammar productions consisting of the ruleset most likely to occur. The performance of our work is evaluated using a comprehensive benchmarking database of on-line symbols.

The aim of writer identification is determining the writer of a piece of handwriting from a set o... more The aim of writer identification is determining the writer of a piece of handwriting from a set of writers. In this paper we present a system for writer identification in old handwritten music scores. Even though an important amount of compositions contains handwritten text in the music scores, the aim of our work is to use only music notation to determine the author. The steps of the system proposed are the following. First of all, the music sheet is preprocessed and normalized for obtaining a single binarized music line, without the staff lines. Afterwards, 100 features are extracted for every music line, which are subsequently used in a k-NN classifier that compares every feature vector with prototypes stored in a database. By applying feature selection and extraction methods on the original feature set, the performance is increased. The proposed method has been tested on a database of old music scores from the 17th to 19th centuries, achieving a recognition rate of about 95%.

One of the major difficulties of handwriting recognition is the variability among symbols because... more One of the major difficulties of handwriting recognition is the variability among symbols because of the different writer styles. In this paper we introduce the boosting of blurred shape models with error correction, which is a robust approach for describing and recognizing handwritten symbols tolerant to this variability. A symbol is described by a probability density function of blurred shape model that encodes the probability of pixel densities of image regions. Then, to learn the most distinctive features among symbol classes, boosting techniques are used to maximize the separability among the blurred shape models. Finally, the set of binary boosting classifiers is embedded in the framework of Error Correcting Output Codes (ECOC). Our approach has been evaluated in two benchmarking scenarios consisting of handwritten symbols. Compared with state-of-the-art descriptors, our method shows higher tolerance to the irregular deformations induced by handwritten strokes.

The recent advances in sketch-based applications and digital-pen protocols make visual languages ... more The recent advances in sketch-based applications and digital-pen protocols make visual languages useful tools for Human Computer Interaction. Graphical symbols are the core elements of a sketch and, hence a visual language. Thus, symbol recognition approaches are the basis for visual language parsing. In this paper we propose an adjacency grammar to represent graphical symbols in a sketchy framework. Adjacency grammars represent the visual syntax in terms of adjacency relations between primitives. Graphical symbols may be either diagram components or gestures. An on-line parsing method is also proposed. The performance of the recognition is evaluated using a benchmarking database of 5000 on-line symbols. Finally, an application framework for sketching architectural floor plans is described.

This article describes a sketch-based framework for semi-automatic annotation of historical docum... more This article describes a sketch-based framework for semi-automatic annotation of historical document collections. It is motivated by the fact that fully automatic methods, while helpful for extracting metadata from large collections, have two main drawbacks in a real-world application: (i) they are error-prone and manual intervention is always required, and (ii) they only capture a subset of all the knowledge in the document base. Therefore, we have developed a practical framework for allowing experts to extract knowledge from document collections in a sketchbased scenario. The main possibilities of the proposed framework are: (a) browsing the collection efficiently, (b) providing gestures for metadata input, (c) supporting handwritten notes and (d) providing gestures for launching automatic extraction processes such as OCR or word spotting.

Symbol recognition is a well-known challenge in the field of graphics recognition. A symbol can b... more Symbol recognition is a well-known challenge in the field of graphics recognition. A symbol can be defined as a structure within a document that has a particular meaning in the context of the application. Due to their representational power, graph structures are usually used to represent line drawings images. Thus, a number of graph comparison approaches are required to answer whether a known symbol appears in a document and under which degree of confidence. In this paper we propose two strategies to recognize symbols depending on the type of their substructures. For those symbols that can be defined by a prototype pattern, we propose a graph isomorphism approach. On the other hand, for those structures consisting of repetitive patterns, we propose a syntactic approach based on graph grammars.

Syntactic approaches on structural symbol recognition are characterized by defining symbols using... more Syntactic approaches on structural symbol recognition are characterized by defining symbols using a grammar. Following the grammar productions a parser is constructed to recognize symbols: given an input, the parser detects whether it belongs to the language generated by the grammar, recognizing the symbol, or not. In this paper, we describe a parsing methodology to recognize a set of symbols represented by an adjacency grammar. An adjacency grammar is a grammar that describes a symbol in terms of the primitives that form it and the relations among these primitives. These relations are called constraints, which are validated using a defined cost function. The cost function approximates the distortion degree associated to the constraint. When a symbol has been recognized the cost associated to the symbol is like a similarity value. The evaluation of the method has been realized from a qualitative point of view, asking some users to draw some sketches. From a quantitative point of view a benchmarking database of sketched symbols has been used.

The analysis of historical document images is not only interesting for the preservation of histor... more The analysis of historical document images is not only interesting for the preservation of historical heritage but also for the extraction of semantic knowledge. In this paper we present a word spotting approach to find keyword images in digital archives. Detected words allow to construct metadata on document contents for indexing and retrieval purposes. Instead of using OCR based approches that would require accurate segmentation and high image quality, we propose a shape recognition method based on the well-known shape context descriptor. Our method is proven to be robust under hightly distorted and noisy document images, a usual drawback in old document analysis. It has been used in a real application scenario, the Collection of Border Records of the Girona Archive. In particular, spotted keywords are used to extract knowledge on personal data of people referred in the documents.

Optical Music Recognition consists in the identification of music information from images of scor... more Optical Music Recognition consists in the identification of music information from images of scores. In this paper, we propose a method for the early stages of the recognition: segmentation of staff lines and graphical primitives in handwritten scores. After introducing our work with modern musical scores (where projections and Hough Transform are effectively used), an approach to deal with ancient handwritten scores is exposed. The recognition of such these old scores is more difficult due to paper degradation and the lack of a standard in musical notation. Our method has been tested with several scores of 19th century with high performance rates.

In this paper we present a colour segmentation method based on a normalized colour naming algorit... more In this paper we present a colour segmentation method based on a normalized colour naming algorithm which removes the effects of the varying conditions due to changes in scene illuminant. Images labelled with the colour name and intensity of small regions are further processed by a region growing step providing a sound segmentation. The method has been tested on a large set of images we get from a surveillance system, whose goal is the automatic retrieval of people from an image database using their appearance description. It is given in terms of placement of the colour regions in clothes. Finally, a quantitative measurement to evaluate the performance of the algorithm has been defined.

Lung, 2006

Pattern Analysis and Applications, 2010

Pattern Recognition Letters, 2009