Utilization of Optical Character Recognition (OCR) in the Development of a Number System Converter Application

Objectives: This study focused on integrating Optical Character Recognition (OCR) in the development of a Number System Converter Android application. The primary goal of the study was to read the written or printed text either in decimal, binary, hexadecimal and octal and convert it to the desired number system. Methods: The Rapid Application Development methodology was employed in the development of the system. Findings: The expert testing resulted in a grand mean of 4.77, interpreted as strongly acceptable. This means that the expert evaluators were able to use the system with ease and was easily able to convert text to the desired number system successfully. Applications/Improvements: The successful development of the application will immensely help students to easily convert a value from one number system to another.


Introduction
Information technology has revolutionized and transformed how people deal with the economy, business, politics, education, religion, and many other fields. Every day more and more innovations are developed with the primary aim to provide convenience with the way people learn and communicate 1 . One of the major breakthroughs that made a huge impact in many fields is the development of mobile and wireless technology. One of the fields that have been greatly impacted by mobile technology is the academe. This started the shift of many educational institutions from a traditional setting to a mobile learning setting.
One of the emerging technologies that are being utilized by many mobile application developers is the OCR. OCR is the mechanical or electronic conversion of images of typed, handwritten or printed text into machine-encoded text, whether from a scanned document, a photo of a document, a scene-photo or from subtitle text superimposed on an image 2 . It is widely used as a form of information entry from printed paper data records, whether passport documents, invoices, bank statements, computerized receipts, business cards, mail, printouts of static-data or any suitable documentation.
Meanwhile, one of the most cognitively challenging topics in computer science is the topic on number system conversion. Number system conversion involves converting a value from one base to another other (e.g. decimal to octal, decimal to binaryf5 and vice versa). All number systems are interconvertable, each having a different method to be converted 3 . The long conversion methods become very tedious for beginners making it hard for them to understand and surpass the number system conversion topic 4 .
The mobile technology revolution and the OCR technology have prompted the researcher to develop a mobile application that utilized the capabilities of Optical Character Recognition. The application served as an automated number converter but instead of typing the data the application was integrated with a camera that reads a handwritten text and automatically converts it to the desired number system (either decimal, binary, octal or hexadecimal).

Objectives of the Study
This study primarily aimed to integrate OCR in the development of an Android-based Number System Converter Application. Specifically, this study aimed to:

Software Development Methodology
The development of the system adapted the Rapid Application Development (RAD) model. Rapid application development is a software methodology which provides a faster development and does not compromise the quality of the software 6 .As shown in Figure 1, RAD is divided into three major stages: (1) requirements planning; (2) system design; and (3) system implementation 7,8 .

Requirement Planning Phase
The developers conducted on-site observations and interviews to be familiar with the transactions of the manual system and assess the needs of the users. System requirements were then identified to fit the needs of the users.

RAD Design Workshop Phase
The developers developed module designs and integrate them into a functional system. The module designs and functions were updated depending on the user responses. The process was done over and over until a refined version of the system was developed.

Implementation Phase
The refined versions were then tested by the end-user/ clients for final quality check and final implementation is carried out once the system adheres to the end-users' standards. Figure 2 shows the different capabilities of the application. As shown in the use case diagram, the user can open the application, choose the desired number system and ultimately capture the written value; the application will then automatically convert the captured value and display the final output.

Design Requirements
As shown in Figure 3, the different steps in optical character recognition are as follows: 1. First step in OCR is the Scanning of images and converts it to its corresponding digital value. This process is called text digitization 2. Pre-processing is the application of different algorithms and techniques to make raw data more usable. 3. Noise Detection and Correction is the removal of unwanted data from the converted images and text to ensure that a more usable digital data is captured. 4. Sometimes during the image conversion, digitized data is skewed; this is why skew detection and correction is employed to fix image angles. 5. The primary aim of layout analysis is to divide the raw image into non-text areas and "text lines"-sub-images Vol 12 (16) | April 2019 | www.indjst.org Niel Francis B. Casillano of the original image that each contains a linear arrangement of symbols in the target language. 6. Characters are then divided into arcs, circles, and other geometric figures and are compared to a known character. 7. Finally, neural networks, decision trees, and other algorithms are employed to produce the actual value of the character.

Software Evaluation
The questionnaire that was used to evaluate the system is based on ISO 9126 standards. ISO 9126 serves as a framework or model for providing worldwide acceptable software qualities required for software evaluation. Under this standard, software must possess six main qualities namely: Functionality, Maintainability, Efficiency, Reliability, Portability and Usability. (ISO, 1991). The experts were chosen because of their strong proficiency on the field of mobile computing technology. The experts were faculty members who handled subjects related to android technology, java application development and web applet development.

Results and Discussion
After undergoing the different phases of development and software evaluation under the RAD model, the following results were obtained: Figure 4 shows the homo form for the application. The user can choose whether to convert a value from decimal (base 10) to binary (base 2), octal (base 8), and hexadecimal (base 16).    Table 1 shows that all experts answered yes (n=15, 100%) to all items under the parameter functionality. This means that all components needed by the end-user/client are present in the mobile application. This also means that the system has fully adhered to the standards of ISO in terms of functionality. Table 2 shows a grand mean of 4.91 interpreted as strongly acceptable. All five subparameters were rated as strongly acceptable (understand ability; learn ability, operability, attractiveness, usability compliance). This result implies that the mobile system is indeed usable. The mobile system can easily be manoeuvred and all its components and operations. Its basic design also made the system learnable and user-friendly. The same data also entails that the mobile system adhered to the standards of ISO in terms of Usability.  Table 3 shows a grand mean of 4.63 interpreted as strongly acceptable. All three subparameters were rated as strongly acceptable (Time behavior, Resource Utilization, and Efficiency Compliance). This result implies that the mobile system is efficient in terms of delivering results and handling data. This also means that the mobile system fully adhered to the ISO standards in term of Efficiency.  Table 4 shows us an overall mean of 4.77 interpreted as strongly acceptable. The parameter Functionality, as shown in Table 1, shows that all the needed requirements based on the requirements analysis were all met; this is evident with the full agreement of all expert testers. The major parameters (Efficiency and Usability) were rated by the pool of experts as Strongly Acceptable. The different parameters were ranked based on the weighted mean to determine the best parameter the mobile system has and based on table 4 that best feature the system is its Usability. Although the mobile system was deemed strongly acceptable improvements can be made so that all parameters will have a perfect rating and for the mobile system to be more resistant to errors and bugs. Overall, the system has adhered to the quality standards of ISO 9126.

Conclusions
Based on the System Development and Expert Testing, the following conclusions were drawn: 1. The following features were successfully developed and integrated into the system, as per results reflected on 2. The Expert Testing resulted in a grand mean of 4.77, interpreted as strongly acceptable. This means that the system adhered to all sub-parameters under the major parameters Efficiency and Usability. Furthermore, improvements in the overall design and fault tolerance are recommended to improve the results of the expert testing.