R. P. Neves

A SVM based off-line handwritten digit recognizer

This paper presents an efficient method for handwritten digit recognition. The proposed method makes use of Support Vector Machines (SVM), benefitting from its generalization power. The method presents improved recognition rates when compared to Multi-Layer Perceptron (MLP) classifiers, other SVM classifiers and hybrid classifiers. Experiments and comparisons were done using a digit set extracted from the NIST SD19 digit database. The proposed SVM method achieved higher recognition rates and it outperformed other methods. It is also shown that although using solely SVMs for the task, the new method does not suffer when considering processing time.

A local thresholding algorithm for images of handwritten historical documents

This paper presents a new algorithm to threshold images of historical documents. The thresholding phase is one of the most important phase in document recognition. An error on it could turn impossible the recognition of the document content. The proposed algorithm is divided into three phases. The first is responsible to indentify the main objects of the image. The second phase divides the image into sub-images according to the previous identification. And the latest phase evaluates a local threshold for each sub-image and proceeds with the binarization of each region. Our approach presented good results in images with complex background and it obtained the best performace when compared with other thresholding algorithms based on the measures used in DIBCO 2009.

An efficient way of combining SVMs for handwritten digit recognition

This paper presents a method of combining SVMs (support vector machines) for multiclass problems that ensures a high recognition rate and a short processing time when compared to other classifiers. This hierarchical SVM combination considers the high recognition rate and short processing time as evaluation criteria. The used case study was the handwritten digit recognition problem with promising results.

Thresholding the Courtesy Amount of Brazilian Bank Checks Based on Tsallis Entropy

Automatic bank check processing is a very hard task as it requires efficiency and precision. Banks have been adding different background patterns in the checks to increase its security. For automatic recognition of the contents of the check, the systems need to remove this background as it does not contain any valid information of the check. This can be done by thresholding algorithms which can work in document images removing all the background and letting just the ink colors to remain. In this paper, we present a new algorithm to threshold images of the courtesy amount of Brazilian bank checks. This new algorithm uses the concept of entropy as defined by Tsallis to find the best cut-off value. Experiments have shown that our algorithm is better than several well-known thresholding algorithms.

A new algorithm to threshold the courtesy amount of Brazilian bank checks

This paper describes a new algorithm for thresholding the courtesy amount of Brazilian bank checks. These images have complex backgrounds which is a problem for an automatic recognition system. The new approach is based on Tsallis entropy to find the best threshold value. Histogram specification is also used for preprocessing some images. The bi-level images are analyzed through several quantitative measures and it achieved the best results when compared with the images produced by other classical thresholding algorithms.

An adaptive thresholding algorithm based on edge detection and morphological operations for document images

This paper presents a new algorithm to threshold document images. The proposed algorithm deal with complex background images, illumination and aspect variants, back-to-front interference, variation of brightness and different positioned shadows. The algorithm have two phases. The first one uses edge detection and morphological operations to identify the text on the image. The second phase uses the positions of the text to define the threshold value in an adaptive process. Our approach presents promising results in images with complex background released from the Document Image Binarization Contest (DIBCO) when compared with other literature and competition thresholding algorithms.