A New Hybrid Algorithm for Telugu Word Retrieval and Recognition

Abstract

Due to their many applications the optical character recognition (OCR) systems have been developed even for scripts like Telugu. Due to the huge number of symbols utilization, identifying the Telugu words are very much complicated. Pre-computed symbol features have been stored by these types of systems to be recognized or to retrieve in a database. Hence, searching of Telugu script from the database is a challenging task due to the complication in finding the features of the Telugu word images or scripts. Here, we had implemented novel Telugu script recognition and retrieval based on the extraction of texture properties features using iterative partitioned clustering (IPC) for classification of word images. In addition, the statistical feature extraction and similarity matching performance is further improved that measures the similarity between trained and test templates.For testing purpose, we utilized noisy, corrupted and occlusion scanned documents as a query input word images, also considered multi conjunct vowel consonant clustered word images. Our extensive simulation analysis shown that the proposed methodology finds most relevant word images from database even under such conditions. Our proposed scheme has performed superior to the conventional approaches presented in the literature in terms of mean Average Precision (mAP) and mean Average Recall (mAR).

Telugu Word Image Retrieval using Deep learning Convolutional Neural Networks

“…As shown in Figure 5, different kind of Telugu word images like occlusion affected, missing segment, noisy effected, random distortion and missing segment with random distorted images are considered as a query word images. Assessment of proposed TWIR system using DL-CNN is done by computing mean average precision (mAP) and mean average recall (mAR) and compared with the conventional TWIR systems like SIFT-BoVW [14], HMM-C [16], SURF-BoVW [17], GLCM-IPC [18], HWNET v2 [19] and SDM-NSCT [21]. As discussed earlier, simulation analysis is done with several kind of Telugu word images and obtained enhanced mAP and mAR even when the query word images had a kind of unwanted information which might be introduced automatically while acquiring them or manually by a human or even by a printing machine during the scanning procedure.…”

“…A further enhancement of TWIR system is achieved in [18], where the authors employed gray level cooccurrence matrix with iterative partitioned clustering (GLCM-IPC) approach. Additionally, image statistics also computed for more accuracy.…”