CNN Based Approach for Offline Text Detection
Anju Bala1, Surinder Kaur2, Gurpreet Singh3

1Anju Bala, Assistant Professor in Chandigarh University, Computer Science Engineering, Gharuan (Punjab), India.
2Surinder Kaur, Assistant Professor, Department of Computer Science & Engineering, Chandigarh University, Gharuan (Punjab), India.
3Gurpreet Singh, Assistant Professor, Department of Computer Science & Engineering, Punjab Technical University, Jalandhar (Punjab), India.

Manuscript received on 18 June 2019 | Revised Manuscript received on 25 June 2019 | Manuscript published on 30 June 2019 | PP: 2582-2584 | Volume-8 Issue-5, June 2019 | Retrieval Number: E7603068519/19©BEIESP
Open Access | Ethics and Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (

Abstract: Offline text detection is a research area where the scanned images containing printed text have been considered as input and then processed to convert the text into digital form that can be understandable and further manipulated by the computer system. In these applications the major difficulty is to segment the area containing text and then after segmentation extracting the meaning of symbols present in the segmented area. The later step is known as classification of characters and words. In this paper a system has been proposed and implemented based on convolution neural network (CNN) classifier to extract the handwritten text from the scanned images. The targeted text has been considered to be written with the help of Roman script means English language. Classifier performed well and produced results at character level up to 97% perfection.
Keywords: Roman, Cnn, English, Text Detection, Classifier

Scope of the Article: Text Mining