Improving the quality of degraded images is a key problem in image processing, but the breadth of the problem leads to domain-specific approaches for tasks such as super-resolution and compression artifact removal. Recent approaches have shown that a general approach is possible by learning application-specific models from examples; however, learning models sophisticated enough to generate high-quality images is computationally expensive, and so specific per-application or per-dataset models are impractical. To solve this problem, we present an efficient semi-local approximation scheme to large-scale Gaussian processes. This allows efficient learning of task-specific image enhancements from example images without reducing quality. As such, our algorithm can be easily customized to specific applications and datasets, and we show the efficiency and effectiveness of our approach across five domains: single-image super-resolution for scene, human face, and text images, and artifact removal in JPEG- and JPEG 2000-encoded images.

Scene text recognition is the problem of recognizing text in natural scene images taken in unconstrained manner. Many approaches have been proposed, but most of them utilize character models only in character recognition phase, the last stage of the process. If characteristic of the target character shape is utilized earlier for the text detection and extraction, text localization would be much easier and therefore, text recognition would be more robust. In this paper, we propose a novel scene text recognition approach which fully utilizes models of target characters from the beginning to the end of the recognition process. Each of target character set is modeled with part-based object model called Implicit Shape Model (ISM) to achieve robustness for partial degradation of text objects. Towards this end, we trained a Hough forest which localizes and aggregates character parts to detect characters candidates from an image. The detected character candidates are verified by organizing the most plausible text lines using dynamic programming. As concrete character models are utilized throughout the process, even extremely deformed texts are detected and recognized, which are hardly detected with previous approaches.

This paper presents an overview of the second Competition on Recognition of Online Handwritten Mathematical Expressions, CROHME 2012. The objective of the contest is to identify current advances in mathematical expression recognition using common evaluation performance measures and datasets. This paper describes the contest details including the evaluation measures used as well as the performance of the 7 submitted systems along with a short description of each system. Progress as compared to the 1st version of CROHME is also documented.

Features and relationships based on character color, edge, stroke and context plays a role for text extraction in natural scene images, but any single feature or relationship is not enough to do the job. This paper presents a novel approach for combining features and relationships within the Conditional Random Field (CRF) framework. By a simple homogeneity measure, an input image is over segmented into perceptually meaningful super pixels and then the text extraction task is formulated as a problem of super pixel labeling. Such a formulation allows us to achieve parameter learning from training images and probabilistic inferences by combining all the features and relationships of the input image. The proposed method shows high performance, in terms of quality, on both the KAIST scene text DB and the ICDAR 2003 DB.

A competition on recognition of online handwritten mathematical expressions is organized. Recognition of mathematical expressions has been an attractive problem for the pattern recognition community because of the presence of enormous uncertainties and ambiguities as encountered during parsing of the two-dimensional structure of expressions. The goal of this competition is to bring out a state of the art for the related research. Three labs come together to organize the event and six other research groups participated the competition. The competition defines a standard format for presenting information, provides a training set of 921 expressions and supplies the underlying grammar for understanding the content of the training data. Participants were invited to submit their recognizers which were tested with a new set of 348 expressions. Systems are evaluated based on four different aspects of the recognition problem. However, the final rating of the systems is done based on their correct expression recognition accuracies. The best expression level recognition accuracy (on the test data) shown by the competing systems is 19.83% whereas a baseline system developed by one of the organizing groups reports an accuracy 22.41% on the same data set.

In handwritten mathematical expressions (ME), understanding the general structure of an ME is often easier than resolving local ambiguities. For instance, identifying a key operator in terms of its spatial relationship with its subordinates is relatively easier than resolving the ambiguities of single symbol identity and local spatial relationships. In addition, decisions related to key operators often occur close to the top (root) of the parse tree, while local decisions take place at the bottom of it. Based on these observations, we propose an incremental search framework in which a parse tree is expanded by tentatively selecting the key operators of an expression. The goodness of the selection is defined by the likelihood of key symbol, the goodness of the sub expressions, and their spatial relationships. In this framework, ambiguous local parts are processed after tentative decisions have been made at the global level. To handle explosiveness of key operator selection, an admissible heuristic function is defined based on the direct relationship of the key operator with the symbols at the bottom level. An experimental evaluation shows that our system is promising. Using it a robust interpretation can be made by utilizing global information and an interpretation can be reached quickly by the admissible heuristic function.

In this paper, we propose a framework for isolating text regions from natural scene images. The main algorithm has two functions: it generates text region candidates, and it verifies of the label of the candidates (text or non-text). The text region candidates are generated through a modified K-means clustering algorithm, which references texture features, edge information and color information. The candidate labels are then verified in a global sense by the Markov Random Field model where collinearity weight is added as long as most texts are aligned. The proposed method achieves reasonable accuracy for text extraction from moderately difficult examples from the ICDAR 2003 database.

Robust extraction of text from scene images is essential for successful scene text recognition. Scene images usually have non-uniform illumination, complex background, and existence of text-like objects. The common assumption of a homogeneous text region on a nearly uniform background cannot be maintained in real applications. We proposed a text extraction method that utilizes user's hint on the location of the text within the image. A resizable square rim in the viewfinder of the mobile camera, referred to here as a 'focus', is the interface used to help the user indicate the target text. With the hint from the focus, the color of the target text is easily estimated by clustering colors only within the focused section. Image binarization with the estimated color is performed to extract connected components. After obtaining the text region within the focused section, the text region is expanded iteratively by searching neighboring regions with the updated text color. Such an iterative method would prevent the problem of one text region being separated into more than one component due to non-uniform illumination and reflection. A text verification process is conducted on the extracted components to determine the true text region. It is demonstrated that the proposed method achieved high accuracy of text extraction for moderately difficult examples from the ICDAR 2003 database.

This paper presents a rule-based approach that utilizes some types of contextual information to improve the accuracy of handwritten mathematical expression(ME) recognition. Mining context from corpus is not practical for ME recognition due to the complexity originated from 2-D nature of MEs. For practicality, we identify typical types of consistencies that are often found in customary usage and general patterns in MEs. We aim to increase these consistencies in recognition results by correcting symbol labels and/or spatial relations among symbols. Such consistencies are easily encoded as condition-action pairs. Preliminary interpretations generated by the base recognizer are reordered by increasing or decreasing scores by the rules. Although our approach is not complete, it easily implements even global context among distant symbols. Experimental results show that our approach is useful to increase the accuracy of handwritten ME recognition.

This paper presents a text image super resolution algorithm based on total variation (TV). Text images typically consist of slim strokes on background. Thus, there are three different local characteristics as homogeneous, directed and complex on text image. Homogeneous region corresponds to background and directed means the region with dominant stroke direction and remaining is complex region. We proposed higher order smoothing on homogeneous region and anisotropic regularization on directed region which encodes the preference of edge direction by smoothing along preferred direction only. Required regularization terms are combined in proposed anisotropic TV functional and controlled by relating parameters. We calculated relating parameters byutilizing structure tensor field. Also to reduce the computational cost, we previously estimated nonchanging pixels and exclude them from calculation for speed up. Experiments shown that, proposed method performs better with low computational cost than general purpose TV on text image.

This paper presents a prior model for text image super-resolution in the Bayesian framework. In contrast to generic image super-resolution task, super-resolution of text images can be benefited from strong prior knowledge of the image class: firstly, low-resolution images are assumed to be generated from a high-resolution image by a sort of degradation which can be grasped through example pairs of the original and the corresponding degradation; secondly, text images are composed of two homogeneous regions, text and background regions. These properties were represented in a Markov random field (MRF) framework. Experiments showed that our model is more appropriate to text image super-resolution than the other prior models.

In structural character recognition, a character is usually viewed as a set of strokes and the spatial relationships between them. Therefore, strokes and their relationships should be properly modeled for effective character representation. For this purpose, we propose a modeling scheme by which strokes as well as relationships are stochastically represented by utilizing the hierarchical characteristics of target characters. A character is defined by a multivariate random variable over the components and its probability distribution is learned from a training data set. To overcome difficulties of the learning due to the high order of the probability distribution (a problem of curse of dimensionality), the probability distribution is factorized and approximated by a set of lower-order probability distributions by applying the idea of relationship decomposition recursively to components and subcomponents. Based on the proposed method, a handwritten Hangul (Korean) character recognition system is developed. Recognition experiments conducted on a public database show the effectiveness of the proposed relationship modeling. The recognition accuracy increased by 5.5 percent in comparison to the most successful system ever reported.

In this paper we propose a Bayesian network framework for explicitly modeling components and their relationships of Korean Hangul characters. A Hangul character is modeled with hierarchical components: a syllable model, grapheme models, stroke models and point models. Each model is constructed with subcomponents and their relationships except a point model, the primitive one, which is represented by a 2D Gaussian for X-Y coordinates of a point instances. Relationships between components are modeled with their positional dependencies. For online handwritten Hangul characters, the proposed system shows higher recognition rates than the HMM system with chain code features: 95.7% vs. 92.9% on average.

In this paper, we propose a new character generationmethod from on-line handwriting recognizers basedon Bayesian networks. On-line handwriting recognizersare trained with handwriting samples from many writers.Then, character shapes are generated from given texts bysearching the most probable input point sequences. SinceBayesian network based classifiers have large number ofparameters for modeling components and their relationships,they generate more natural character shapes thanvarious kinds of hidden Markov models.

In structural character recognition, a character is usually viewed as a set of strokes and the spatial relationships between them. In this paper, we propose a stochastic modeling scheme by which strokes as well as relationships are represented by utilizing the hierarchical characteristics of target characters. Based on the proposed scheme, a handwritten Hangul (Korean) character recognition system is developed. The effectiveness of the proposed scheme is shown through experimental results conducted on a public database.

In this paper, we propose a novel method of building a language model for open-vocabulary Korean word recognition. Due to the complex morphology of Korean, it is inappropriate to use lexicons based on the linguistic entities such as words and morphemes in open-vocabulary domains. Instead, we build the lexicon by collecting variable length character sequences from the raw texts using a dynamic Bayesian network model of the language. In simulated word recognition experiments, the proposed language model could find correct words from lattices of character candidates in 94.3% of cases, increasing the word recognition rates by 20.9%.

The current paper presents a novel texture-based method for detecting texts in images. A support vector machine (SVM) is used to analyze the textural properties of texts. No external texture feature extraction module is used, but rather the intensities of the raw pixels that make up the textural pattern are fed directly to the SVM, which works well even in high-dimensional spaces. Next, text regions are identified by applying a continuously adaptive mean shift algorithm (CAMSHIFT) to the results of the texture analysis. The combination of CAMSHIFT and SVMs produces both robust and efficient text detection, as time-consuming texture analyses for less relevant pixels are restricted, leaving only a small part of the input image to be texture-analyzed.

This paper proposes a statistical character structure modeling method. It represents each stroke by the distribution of the feature points. The character structure is represented by the joint distribution of the component strokes. In the proposed model, the stroke relationship is effectively reflected by the statistical dependency. It can represent all kinds of stroke relationship effectively in a systematic way. Based on the character representation, a stroke neighbor selection method is also proposed. It measures the importance of a stroke relationship by the mutual information among the strokes. With such a measure, the important neighbor relationships are selected by the nth order probability approximation method. The neighbor selection algorithm reduces the complexity significantly because we can reflect only some important relationships instead of all existing relationships. The proposed character modeling method was applied to a handwritten Chinese character recognition system. Applying a model-driven stroke extraction algorithm that cooperates with a selective matching algorithm, the proposed system is better than conventional structural recognition systems in analyzing degraded images. The effectiveness of the proposed methods was visualized by the experiments. The proposed method successfully detected and reflected the stroke relationships that seemed intuitively important. The overall recognition rate was 98.45 percent, which confirms the effectiveness of the proposed methods.

This paper presents a real-time face recognition system. For the system to be real time, no external time-consuming feature extraction method is used, rather the gray-level values of the raw pixels that make up the face pattern are fed directly to the recognizer. In order to absorb the resulting high dimensionality of the input space, support vector machines (SVMs), which are known to work well even in high-dimensional space, are used as the face recognizer. Furthermore, a modified form of polynomial kernel (local correlation kernel) is utilized to take account of prior knowledge about facial structures and is used as the alternative feature extractor. Since SVMs were originally developed for two-class classification, their basic scheme is extended for multiface recognition by adopting one-per-class decomposition. In order to make a final classification from several one-per-class SVM outputs, a neural network (NN) is used as the arbitrator. Experiments with ORL database show a recognition rate of 97.9% and speed of 0.22 seconds per face with 40 classes.

In this paper a framework for off-line handwritten numeral string recognition based on stroke grouping is proposed. In our approach, strokes are aligned into a sequence of strokes and then a segmentation process is performed to partition strokes in the sequence into possible-digits, that is, groups of strokes which may be a digit. As a result of stroke grouping, grouping-hypotheses, which imply possible segmentation, are generated. An input numeral string is recognized by a dynamic programming scheme, in which the best grouping-hypothesis with maximum matching score is chosen. The framework also provides a systematic way of reducing computational complexity by embedding external knowledge into the framework. The experimental results to evaluate the proposed framework are shown.

Structural matching and feature matching are two major streams in handwritten Chinese character recognition. The former is better in absorbing writing variation since it utilizes rich structural information from strokes directly. However, it is worse than the latter in handling noisy images, because of the difficulty in stroke extraction. For the ultimate performance in a recognition system, the structural information should be utilized without suffering from the faults of stroke extraction. This paper proposes a stroke-guided pixel matching method to satisfy the requirements. It uses a pixel-based character model to match images not depending on strokes. In order to utilize structural information, a stroke-guided pixel-grouping algorithm is also proposed. The initial experimental result shows that the proposed approach is promising.

The idea of combining the network of HMMs and the dynamic programming-based search is highly relevant to online handwriting recognition. The word model of HMM network can be systematically constructed by concatenating letter and ligature HMM's while sharing common ones. Character recognition in such a network can be defined as the task of best aligning a given input sequence to the best path in the network. One distinguishing feature of the approach is that letter segmentation is obtained simultaneously with recognition but no extra computation is required.

It is well known that the stochastic approach using the HMM and dynamic programming-based search is particularly suited to the analysis of time series signals including on-line handwriting. The starting point of this research is a network of HMMs which models the whole set of characters. Then it is followed by the assertion that the HMM for the on-line script can be applied to not only on-line character recognition but also to the handwriting synthesis and even pen-trajectory recovery in off-line character images. The solutions to these problems are based on the single network of HMMs and the single principle of DP-based state-observation alignment. Given an observation sequence, the search for the best path in the network corresponds to the recognition. Given a character model, the search for the best observation sequence corresponds to the handwriting generation. The proposed framework has been shown to work nicely through a set of tests on Korean characters.

Most previous studies assumed that classifiers behave independently. Such an assumption degrades and biases classification performance in the case of adding highly dependent classifiers. In order to overcome such a weakness, one should consider combining multiple classifiers in a probabilistic framework using a Bayesian formalism without the assumption. The probabilistic combination of K classifiers needs a (K+1)st-order probability distribution. However, it as well known that the distribution becomes unmanageable when storing and estimating, even for small K. Chow and Liu (1968) as well as Lewis (1959) proposed a product approximation of a high order distribution with a set of only first-order tree dependencies or second-order distributions. However, if a classifier follows more than one classifier, such a first-order dependency will not be suitable to approximate the high order distribution properly. A probabilistic framework is proposed to identify an optimal product set of kth-order dependencies for an approximation of the (K+1)st-order probability distribution where 1/spl les/k/spl les/K, and to combine K decisions by the identified product set using a Bayesian formalism. This framework was experimented and evaluated with a standardized CENPARMI database and showed superior performance when compared to other combination methods.

Proposes some strategies to improve the recognition performance of a feature matching method for handwritten Chinese character recognition (HCCR). Favorable modifications are given to all stages throughout the recognition. In pre-processing, we devised a modified nonlinear normalization algorithm and a connectivity-preserving smoothing algorithm. For feature extraction, an efficient directional decomposition algorithm and a systematic approach to design a blurring mask are presented. Finally, a modified LVQ3 algorithm is applied to optimize the reference vectors for classification. The integrated effect of these strategies significantly improves the recognition performance. Recognition results on the large-vocabulary databases ETL8B2 and ETL9B are promising.

Online recognition of cursive words is a difficult task owing to variable shape and ambiguous letter boundaries. The approach proposed is based on hidden Markov modeling of letters and inter-letter patterns called ligatures occurring in cursive script. For each of the letters and the ligatures we create one HMM that models temporal and spatial variability of handwriting. By networking the two kinds of HMMs, we can design a network model for all words or composite characters. The network incorporates the knowledge sources of grammatical and structural constraints so that it can better capture the characteristics of handwriting. Given the network, the problem of recognition is formulated into that of finding the most likely path from the start node to the end node. A dynamic programming-based search for the optimal input-network alignment performs character recognition and letter segmentation simultaneously and efficiently. Experiments on Korean character showed correct recognition of up to 93.3% on unconstrained samples. It has also been compared with several other schemes of HMM-based recognition to characterize the proposed approach.

In this paper, we proposed an approach for segmentation and recognition of unconstrained handwritten English words with character and ligature modeling. Viewing a handwritten word as an alternating sequence of characters and ligatures, a network of circularly interconnected hidden Markov models is constructed to model handwritten English words of indefinite length. Then the recognition problem is regarded as finding the maximal probability path in the network for given input sequence. From the path, optimal segmentation and associated character labels are obtained simultaneously without any explicit segmentation.

An automatic system that reads an image of a logic diagram, digitized by a document scanner, and generates a description of the diagram in terms of logic symbols and their interconnections is proposed. The digitized image is converted into a set of line segments by way of a sequence of picture processing operations, Then symbols and connections are recognized by identifying loops and rectilinear polylines. Graphical description of a set of symbol models is provided as prior knowledge about input diagrams. Experiments show that the system can recognize more than 96% of logic symbols and connections and that an A4 size diagram (210×297 mm ) of average complexity can be processed within 10 seconds on a workstation.

A statistical approach to recognizing on-line cursive Hangul character is proposed. Viewing a handwritten Hangul syllable as an alternating sequence of letters and ligatures, all handwritten legal characters are modeled with a finite state network that is a concatenation of letter and ligature HMMs. Given an input to the network, recognition, which corresponds to finding the most likely path, is performed using the dynamic programming technique. Experiments have shown that letter boundary detection as well as handwriting variability resolution is achieved with good results.>

In this paper, a model-based scheme for recognizing hand-drawn symbols in schematic diagrams using attributed graph (AG) matching in the absence of any information concerning their pose (translation, rotation and scale) is described. The process of AG matching proceeds as follows. First, an observed AG (AG O ) is constructed from single-pixel-width line-representation of an observed symbol. Second, the pose of the AG O is estimated in terms of translation, rotation and scale with respect to the model AGs (AG M s ). The search space is effectively pruned by introducing the concept of control vertex and applying geometrical constraints in an early stage. In this step, a small number of candidate AG M s are selected. Third, correspondences between components of the observed AG after normalization (AG O N ) and those of the AG M s are found for the given poses. Fourth, distance measures between the AG O N and the AG M s are calculated, based upon the correspondences. Finally, the AG O N is classified as the AG M with the minimum distance. Experimental results for hand-drawn symbols with and without templates show that using AG matching is very efficient and successful for translation-, rotation- and scale-invariant recognition of hand-drawn symbols in schematic diagrams.

A basic minimum-square-error (MSE) transform for point pattern matching and a fast computational method for its implementation are described. Currently, the method is being used successfully in a number of research projects. A computational analysis shows that it is faster than that of F.C.A. Groen et al.>