Deformed character recognition using convolutional neural networks

Abstract

Realization of high accuracies towards south Indian character recognition is one the truly interesting research challenge. In this paper, our investigation is focused on recognition of one of the most widely used south Indian script called Kannada. In particular, the proposed exper-iment is subject towards the recognition of degraded character images which are extracted from the ancient Kannada poetry documents and also on the handwritten character images that are collected from various unconstrained environments. The character images in the degraded documents are slightly blurry as a result of which character image is imposed by a kind of broken and messy appearances, this particular aspect leads to various conflicting behaviors of the recognition algorithm which in turn reduces the accuracy of recognition. The training of degraded patterns of character image samples are carried out by using one of the deep convolution neural networks known as Alex net.The performance evaluation of this experimentation is subject towards the handwritten datasets gathered synthetically from users of age groups between 18-21, 22-25 and 26-30 and also printed datasets which are extracted from ancient document images of Kannada poetry/literature. The datasets are comprised of around 497 classes. 428 classes include consonants, vowels, simple compound characters and complex com-pound characters. Each base character combined with consonant/vowel modifiers in handwritten text with overlapping/touching diacritics are assumed as a separate class in Kannada script for our experimentation. However, for those compound characters that are non-overlapping/touching are still considered as individual classes for which the semantic analysis is carried out during the post processing stage of OCR. It is observed that the performance of the Alex net in classification of printed character samples is reported as 91.3% and with reference to handwritten text, and accuracy of 92% is recorded.

An efficient and improved scheme for handwritten digit recognition based on convolutional neural network

“…This study differs from other existed work because it reveals the effectiveness of CNN in terms of high accuracy and low computational time to classify handwritten digits [36][37][38][39]. With extensive literature survey, it comes to know that accuracy of various reported CNN models is not up to the mark as mentioned in Table 3.…”

“…Their algorithm produced maximum 75% of accuracy. Shobha Rani [39] reported work on recognition of one of the most widely used South Indian script called Kannada. The training of character image samples is carried out by using one of the deep convolution neural networks.…”

Banana Grading Using Deep Learning Model

“…A CNN uses individual components, each of which is referred to as a filter extracted from the original image, to identify character patterns in photographs. [13]. Compare their strategy to those of trained systems like MobileNetV1, InceptionV3, and DenseNet121.…”

Mechanical Properties of Sand-Cement Mortar with Different Water/Cement Ratio under Thermal Gradient Conditions

“…This results in excessive moistening of the latter and further degradation of their properties. Nonuniform temperature fields throughout the cross-section of the structure can have a negative effect on the strength gain of concrete [9][10][11][12][13][14][15]. Cracking caused by the temperature gradient can provide a loss in the structural integrity and a reduction in the service life of in-situ reinforced concrete structures [16][17][18].…”