Savitha Attigeri

Bio: Savitha Attigeri is an academic researcher. The author has contributed to research in topics: Pattern recognition (psychology) & Character (mathematics). The author has an hindex of 1, co-authored 1 publications receiving 11 citations.

Neural Network based Handwritten Character Recognition system

Abstract: Handwritten character recognition has been one of the active and challenging research areas in the field of image processing and pattern recognition. It has numerous applications which include, reading aid for blind, bank cheques and conversion of any hand written document into structural text form. In this paper an attempt is made to recognize handwritten characters for English alphabets without feature extraction using multilayer Feed Forward neural network. Each character data set contains 26 alphabets. Fifty different character data sets are used for training the neural network. The trained network is used for classification and recognition. In the proposed system, each character is resized into 30x20 pixels, which is directly subjected to training. That is, each resized character has 600 pixels and these pixels are taken as features for training the neural network. The results show that the proposed system yields good recognition rates which are comparable to that of feature extraction based schemes for handwritten character recognition

Handwritten Character Recognition Using Deep-Learning

Abstract: In this paper we present an innovative method for offline handwritten character detection using deep neural networks. In today world it has become easier to train deep neural networks because of availability of huge amount of data and various Algorithmic innovations which are taking place. Now-a-days the amount of computational power needed to train a neural network has increased due to the availability of GPU's and other cloud based services like Google Cloud platform and Amazon Web Services which provide resources to train a Neural network on the cloud. We have designed a image segmentation based Handwritten character recognition system. In our system we have made use of OpenCV for performing Image processing and have used Tensorflow for training a the neural Network. We have developed this system using python programming language.

A Review on Handwritten Character Recognition Methods and Techniques

Abstract: In this paper a detailed review of Handwritten Character Recognition is presented. Some features of human beings are unique to individuals like iris, fingerprint, DNA etc. Handwriting is one such feature which is different for each human being and it has been proven scientifically. In Handwritten Character Recognition (HCR) the task is to identify the characters written by humans and converting it into digital text. HCR is a field where plenty of research has been done but still there is scope in terms of improving the accuracy and efficiency. Digitizing manually written text is very useful in today’s world as it makes information readily available anywhere and anytime. Digitized text can be used for commercial purposes and it is more safe and environment friendly as compared to manual text. This review paper will shed some light on various methodologies used till now in this field along with their advantages, limitations and accuracy rate.

Pattern Recognition-Recognition of Handwritten Document Using Convolutional Neural Networks

Abstract: Handwritten documents recognition is a challenging task in the field of pattern recognition. It has an array of applications wherein recognition of words, alphabets, digits and other characters are the mandate. This review article mainly focuses review on Convolutional Neural Network (CNN) based handwritten documents recognition system. Basically, the handwritten recognition is divided into two different types: online and offline recognition. The difficulty of this system is dealing with huge variety of handwritten styles written by different writers. The system wants to recognize and identify such characters in effective manner. The scope of this review paper is to represent the merits and limitations of different techniques used in the development of recognition system. This paper definitely helps the researcher to get new idea to develop a new technique with good environment and architecture to propose more accuracy and less error rate. The major bottlenecks in this system are the issues of recognizing unconstrained handwritings like cursive, block, and tilt that cause huge variation in writing styles, the overlapping and the interconnections between characters. These systems are designed to assure high accuracy and reliability. This article summarizes different strategies in handwritten recognition system, which may help the researchers interested in this area to spot the research gap. The Future work of this paper is to implement a robust technique providing more accuracy and less error rate.

Human Gait Kinematic Estimation based on Joint Data Acquisition and Analysis from IMU and Depth-Sensing Camera

Abstract: Gait analysis is very important on applications of clinical rehabilitation of patients of stroke or spinal cord injuries. In this paper, the acquisition of spatiotemporal and angular parameters are obtained by using a Depth Sensing camera, Inertial Measurement Units (IMUs), and Vicon Motion Capture systems (Vicon MoCap) which is considered as the gold standard in gait data acquisition, simultaneously. However, Vicon MoCap is costly and is not easily accessible in the Philippines. This paper proposes a method for gait analysis through the development of an integrated system for gait data processing, consisting of Depth Sensing Camera and IMU sensors as an alternative to Vicon MoCap. A deep learning algorithm called Bayesian Regularization Artificial Neural Network (BRANN) is used to synthesize the data from Depth Sensing Camera and IMU due to its capabilities in complex and non-linear problems with considerable time. The output of the integrated system, in comparison to the Depth Sensing Camera and IMUs, showed an 85.2647% to 99.5636% reduction for the Depth Sensing Camera and an 82.1329% to 99.6551% reduction for the IMU in the Mean Squared Error. It also showed an increase in the number significant parameters using hypothesis test.

Vietnamese handwritten character recognition using convolutional neural network

Abstract: Handwriting recognition is one of the core applications of computer vision for real-word problems and it has been gaining more interest because of the progression in this field. This paper presents an efficient model for Vietnamese handwriting character recognition by Convolutional Neural Networks (CNNs) – a kind of deep neural network model can achieve high performance on hard recognition tasks. The proposed architecture of the CNN network for Vietnamese handwriting character recognition consists of five hidden layers in which the first 3 layers are convolutional layers and the last 2 layers are fully-connected layers. Overfitting problem is also minimized by using dropout techniques with the reasonable drop rate. The experimental results show that our model achieves approximately 97% accuracy.