Analyze phase

The Analysis phase is the most important phase in the ADDIE model. The analysis phase involves identification of the problems instructional as well as the objectives, and focuses on the environment, knowledge and skills. This level was discussed in Chapter 2 and Chapter 4. The review on automated methods for bone age assessment (BAA) in Chapter 2 highlighted the fundamental problems in image processing techniques and segmentation of the region of interest (ROI) in the hand images. Furthermore, it highlighted that the role of the other bones in young adults, as well as in handling forensic cases where defects and other abnormalities and injuries on the hand are presented in the current automated BAA system were ignored. This is the reason for developing a novel automated BAA system that uses a new image processing approach for the hand and clavicle X-ray images.

The study was carried out to identify the factors that can contribute to the success of developing an automated system for bone age assessment. Hence, a questionnaires survey and interviews were used for data collection.

The respondents were radiologists and specialists from the Faculty of Medicine, University of Malaya. They are considered to be experienced and knowledgeable people in BAA. Based on the outcome of the questionnaire survey and interviews the first research objective has been achieved. The questionnaire survey was aimed at identifying the variables and factors that can affect bone age assessment, from the perspectives of medical specialists. Table 4.9 shows that a high percentage of respondents said that gender, noise in images, and the races of patients are the variables that can affect the assessment of bone age. Hence, this system will consider the effect of these variables in the developments of the system to obtain more accurate results of bone age.

From the results of interviews it can be concluded that automated system for BAA is highly desired by the specialists in UMMC. The results of the interview provide a better

understanding of the problems that doctors face while using the GP atlas for bone age assessment. All respondents agreed on the need to have an automated method for BAA to eliminate observational variability and to reduce the time spent on assessment of bone age. Furthermore, an automated approach for BAA would reduce the time spent and increase the reliability of the final estimation.

The basic target aim in this project is to produce a system that is capable for solving of BAA problems in both the hospital and forensic center, based on the user requirements. The user requirements are to serve as the mandate of reference for the design, development and implementation of a system. The goal of the user requirement is to provide a clear understanding of requirements as an early input to development system (Xie & Meng, 2011). The main requirements are identified from the data collection in Chapter 4 as they are listed in Table 5.1:

Data Collection Method

User Requirements

Survey Questionnaire

 Speed up the process of BAA  Increase the accuracy of BAA

 Eliminate the observation variability

 Using hand and clavicle bone to solve the problem of incomplete or noisy images

 Consider the factors that affect BAA  Record the age for further reference

The findings of collected data address to achieve the requirements in the study and what radiologists and specialists expect from an automated BAA system that will replace the manual methods. Table 5.1 shows, it is expected that our automated system for BAA need to be faster, and assess more accurate results. It is time-consuming for radiologist and doctor to prepare the reports on BAA, hence they require in decreasing the time of BAA in the system. Furthermore, the system that can save the energy and eliminate the human observation would be preferred by the users. Assessment of bone age using a fully automated system removes the need to have people involved and it is eliminates the subjectivity of the estimation.

To design and develop a success information system (IS), it is necessary to identify and consider the end-user’ requirements. Using hand and the clavicle bones could address the missing data or noisy images problem in our system. The implementation of a BAA system based on the hand and clavicle bones can overcome the problems of conducting BAA on patients with hand injuries.

Interview

 Use automated system for BAA  Improve the accuracy in BAA  Decrease the time in assessment  Easy to use

 Have a web based application

Observational Study

 Using Hand images for BAA  Improve the accuracy in assessment  Remove the human observation

The need to introduce an automated BAA system to replace the current manual method is to solve the problem of the repeated work. The BAA system should assess accurate results, and also should store the information more effectively to satisfy the end-user. The following sections explain how all the user requirements are applied in the designing, development and implementation of BAA system.