A Comparison of Feature and Pixel-Based Methods for Recognizing Handwritten Bangla Digits

Abstract

Abstract-We propose a novel handwritten character recognition method for isolated handwritten Bangla digits. A feature is introduced for such patterns, the contour angular technique. It is compared to other methods, such as the hotspot feature, the gray-level normalized character image and a basic lowresolution pixel-based method. One of the goals of this study is to explore performance differences between dedicated feature methods and the pixel-based methods. The four methods are compared with support vector machine (SVM) classifiers on the collection of handwritten Bangla digit images. The results show that the fast contour angular technique outperforms the other techniques when not very many training examples are used. The fast contour angular technique captures aspects of curvature of the handwritten image and results in much faster character classification than the gray pixel-based method. Still, this feature obtains a similar recognition compared to the gray pixel-based method when a large training set is used. In order to investigate further whether the different feature methods represent complementary aspects of shape, the effect of majority voting is explored. The results indicate that the majority voting method achieves the best recognition performance on this dataset.

Evaluating automatically parallelized versions of the support vector machine

“…For instance, the curly tails in Bangla characters makes the definition of a stable bounding box problematic. The best previous results on this dataset were obtained with an ensemble ML technique, where an accuracy of 96.8% was obtained [65].…”

Recognition of handwritten characters using local gradient feature descriptors

“…In Surinta et al (2013), the authors presented the unweighted majority voting method (UMV), which combines different SVM classifiers with the Because the HOG descriptor and siftD with the kNN method provide higher accuracies than the more complex method used in Surinta et al (2013), these results demonstrate the effectiveness of the proposed local gradient feature descriptors. We show the obtained results with the SVM classifier with the RBF kernel on the handwritten character datasets in Table 4.…”

“…In previous studies, the raw image (IMG) method, which directly copies the intensities of the pixels of the ink trace (Surinta et al, 2013), has often been used as the feature extraction method. It extracts a high dimensional feature vector that depends on the size of the input image.…”

A Path Planning for Line Segmentation of Handwritten Documents

“…Such a technique takes into account the diversity of document images, texts, images, mixtures of texts and images, line drawings, and noisy or degraded document images. Bulacu et al [2] and Surinta et al [10] use Otsu's algorithm, a global binarization technique, in their work. Otsu's algorithm uses one threshold value to process an entire document image.…”