Indian Journal of Science and Technology
DOI: 10.17485/IJST/v15i41.1421
Year: 2022, Volume: 15, Issue: 41, Pages: 2121-2128
Original Article
Amit Prakash Sen1*, Nirmal Kumar Rout2, Tuhinansu Pradhan3, Amrit Mukherjee4
1Assistant Professor, School of Engineering, Arka Jain University, Jharkhand, India
2Professor, School of Electronics, KIIT University, Bhubaneswar, India
3Assistant Professor,, Siksha ‘O’ Anusandhan Deemed to be University, Bhubaneswar, Odisha, India
4Assistant Professor, Dept. of Computer Science, University of South Bohemia, Ceske Budejovice, Czech Republic
*Corresponding Author
Email: [email protected]
Received Date:07 September 2022, Accepted Date:22 September 2022, Published Date:02 November 2022
Objectives: To propose a model which will pre-process the dataset for the removal of any noise before the training of the network. Methods: Reported literature does not focus on the pre-processing of the dataset before the training of the network. A noise removal scheme called Probabilistic Decision Based Adaptive Improved Trimmed Median Filter (PDAITMF) is implemented as a pre-processing tool before the developed model. The PDAITMF de-noises the dataset. Findings: This supports an effective learning process by the model. The model is trained, validated, and tested with the respective dataset. Accuracy of 0.9401 is achieved without the implementation of PDAITMF, while an accuracy of 0.9841 is achieved when the model uses the dataset processed by PDAITMF. Synchronization is also established between the training and validation graph which seems to be missing when the model uses the dataset without processing through PDAITMF. Novelty: A sharp improvement in accuracy is noted which establishes the effectiveness of the noise removal scheme before the Deep Learning model. The technique may be used to improve the detection accuracy of other acute diseases.
Keywords: Deep Learning; Transfer Learning; COVID19; Noise removal; COVID 19 detection
© 2022 Sen et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Published By Indian Society for Education and Environment (iSee)
Subscribe now for latest articles and news.