Author(s):
1. Samar Abbas Mangi:
Institute Of Computer Science, Shah Abdul Latif University Khairpur Mirs Sindh, Pakistan
2. Samina Rajper:
Institute Of Computer Science, Shah Abdul Latif University Khairpur Mirs Sindh, Pakistan
3. Jamil Ahmed Chandio:
Institute Of Computer Science, Shah Abdul Latif University Khairpur Mirs Sindh, Pakistan
4. Noor Ahmed Shaikh:
Institute Of Computer Science, Shah Abdul Latif University Khairpur Mirs Sindh, Pakistan
Abstract:
Due to high significance of medical images, the medical image analysis has become well recognized research area of computer science and recent advances in computational technology have boost up the process to alter the image, sound and video data collectively referred to as ISV. During the process of ISV capturing to printing, there are fair chances of ISV forgery, which could become leading case of misdiagnosis in medical images such as Digital X-rays (DX), CTs, MRIs , such unnecessary manipulation with medical image may potentially lead to misdiagnosis. Such unauthorized changes to medical images not only effect the accuracy of diagnosis but also effect the cost of treatment. However forgery detection (FD) at early stage would provide additional assistance to doctors to investigate the health condition patients in more effective way. Since the malicious intensions (MI) of human or malfunctioning of DXs would include extra noise into ISVs, which could become of the leading cause of misdiagnosis but the pixel level image analytics could be used to detect such anomalies. This research offers a framework to detect the forgery in DXs and ISVs. This framework includes three key steps: I. Data Preparation, in which the entropy of forgery area be detected by offering significant method. II. Construction of decision Model: that includes in in which support vector machine (SVM) is used to construct the classical model. III. Visualize the Results in which a confusion matrix and other measures are used to detect the forgeries in the DXs. Our proposed approach contributes highest classification accuracy such as 96.90%.
Page(s):
1-5
DOI:
DOI not available
Published:
Journal: University of Sindh Journal of Information and Communication Technology, Volume: 7, Issue: 1, Year: 2023
Keywords:
machine learning
,
supervised learning
,
Classification
,
Xrays
,
Strong Objects
,
Forgery
References:
References are not available for this document.
Citations
Citations are not available for this document.