Postal Address Recognition Research Articles

CNN-RNN BASED HANDWRITTEN TEXT RECOGNITION

At present most of the scripts are handwritten due to the ease of using a pen tip in place of a keyboard, hence errors are common due to illegibility of the human handwriting. To avoid this problem handwriting recognition is essential. Offline handwritten Text recognition (OHTR) has become one of the major areas of research in recent times because of the need to eliminate errors due to misinterpretation of handwritten text and the need for automation to improve efficiency. The application of this system can be seen in fields like handwritten application interpretations, postal address recognition, signature verification, and various others. In this project, offline handwritten Text recognition is performed using Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM) which uses the architecture of Recurrent Neural network (RNN) and connectionist temporal classification (CTC). The neural network is trained and tested using the IAM database containing handwritten English text. The implementation of this work is done using image segmentation-based handwritten text recognition where OpenCV is used for performing image processing and TensorFlow is used for training and text recognition. This whole system is developed using python and the output is displayed in a word file.

Decision fusion for postal address recognition using belief functions

Combining the outputs from several postal address readers (PARs) is a promising approach for improving the performances of mailing address recognition systems. In this paper, this problem is solved using the Transferable Belief Model, an uncertain reasoning framework based on Dempster–Shafer belief functions. Applying this framework to postal address recognition implies defining the frame of discernment (or set of possible answers to the problem under study), converting PAR outputs into belief functions (taking into account additional information such as confidence scores when available), combining the resulting belief functions, and making decisions. All these steps are detailed in this paper. Experimental results demonstrate the effectiveness of this approach as compared to simple combination rules.

Forty years of research in character and document recognition—an industrial perspective

This paper presents an overview on the last 40-years of technical advances in the field of character and document recognition. Representative developments in each decade are described. Then, key technical developments in the specific area of Kanji recognition in Japan are highlighted. The main part of the paper discusses robustness design principles, which have proven to be effective to solve complex problems in postal address recognition. Included are the hypothesis-driven principle, deferred decision/multiple-hypotheses principle, information integration principle, alternative solution principle, and perturbation principle. Finally, future prospects, the ‘long-tail’ phenomena, and promising new applications are discussed.

Recognition technology frontiers

Rapid improvements in hardware and algorithms are reshaping the technological basis of postal address recognition. An increasingly important theme is automation of the engineering process itself, through trainable classifiers, realistic distortion models, and statistical contextual analysis. Relevant basic research at AT&T Bell Laboratories includes VLSI neural networks, algorithmic pattern recognition, computational linguistics, and artificial intelligence. Interdisciplinary application of these has stimulated improvements in handwritten ZIP code recognition, machine-print address recognition, and address block location.