Neural network based optical character recognition system

Dojčinović, Nikola; Mihajlovic, Igor; Joković, Jugoslav; Marković, Vera; Milovanović, Bratislav

doi:10.1109/neurel.2012.6419976

Cited by 5 publications

(2 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Low resolution images typically origin from cameras with low scanning resolution, under 100dpi [7]. Mathematically, lenses designed for focusing on distances where document elements are not clearly visible, captures blurred scanned images, a blur is modeled as a convolution of the blurring kernel, and the original image if blur origins from lenses not focusing, blurring kernel is Gaussian [8]. Blur decreases the slope of edges, spreading them, occluding near objects, connecting objects unconnected on an original, and its removal significantly increases the accuracy of OCR systems, Other types of noise are also treated [9].…”

Section: Optical Character Recognitionmentioning

confidence: 99%

Character recognition using correlation & hamming distance

Shehu

Ashir

Eleyan

2015

2015 23nd Signal Processing and Communications Applications Conference (SIU)

View full text Add to dashboard Cite

Özetçe-Bu bildiride karakter tanıma, korelasyon kullanılarak şablon eşleme yöntemiyle gerçekleştirilmiştir. Yaklaşımımızda, korelasyona ek olarak, korelasyonun işe yaramadığı durumlarda hamming uzaklığı uygulanmaktadır. Bu çalışmada, metin görüntüsü satırlara ve ardından karakterlere temel önişleme teknikleri ile ayrılmaktadır. Bundan sonra, karakter içeren görüntü sınıflandırılmış ve korelasyon yada hamming mesafesi kullanılarak şablon eşleme yöntemi vasıtasıyla metine dönüştürülmüştür. Test aşaması için, toplam 560 karakterli dört metin görüntüsü oluşturulmuştur. Sistem sadece korelasyon için ortalama %72.39 ve korelasyona ek olarak hamming uzaklığı için %94.90'lük tanıma oranlarıyla başarılı çalışmaktadır..Anahtar Kelimelerkarakter tanıma, korelasyon, hamming uzaklığı.Abstract-In this paper, character recognition is carried out using template matching scheme using correlation. In our approach, in addition to correlation, the hamming distance was applied in the scenarios where the correlation fails. In this approach, text image is segmented into lines and then characters by basic pre-processing techniques. After that, the character image is classified and converted into text by template matching scheme using either correlation or hamming distance. For test stage, four text images were created with total of 560 characters. The system works successfully with average recognition rates of 72.39% and 94.90% for correlation only and correlation plus hamming distance, respectively.

show abstract

Section: Optical Character Recognitionmentioning

confidence: 99%

Character recognition using correlation & hamming distance

Shehu

Ashir

Eleyan

2015

2015 23nd Signal Processing and Communications Applications Conference (SIU)

View full text Add to dashboard Cite

show abstract

“…Dojvcinovic et al [22] presented a neural network based classification of characters in which they have considered whole image as a feature. Extraction and segmentation of characters were recognized in this work.…”

Section: Introductionmentioning

confidence: 99%

Recognition of Degraded and Non Degraded Roman Characters Using Different Classifiers

Verma¹,

Agarwal²,

Aggarwal³

2019

ujeee

View full text Add to dashboard Cite

In this paper, we present an efficient and robust technique for the recognition of offline roman characters. The main strategy is to extract statistical and similarity features using a combination of grey level co-occurrence matrix (GLCM) and complementary similarity measure (CSM) method. In this work, the CSM method is used to extract features from binary images and combined with GLCM to boost the accuracy of character recognition. The recognition has been done using four different classifiers i.e. artificial neural network (ANN), Naive Bayes classifier, random forest (RF) and support vector machine (SVM). The standard dataset has been used for experimental work. We have done experiments on the clean and noisy dataset. It achieves the accuracy of 100% for some characters without noise and 94.11% with impulsive noise. A comparison of these four classifiers is recorded with and without a noisy environment. On a clean dataset, the random forest provides the best average recognition accuracy of 84.9% for all characters. On low noise datasets, random forest and artificial neural networks have almost the same recognition accuracy and on high noise datasets, SVM provides the highest recognition accuracy.

show abstract