The Novel Outlier Extraordinariness Sensing and Recreation Operation Established on Adaptive MDBUTMF
In the research community of the outlier extraordinariness sensing and recreation operation, illustrational units are algebraically manifested as normal and abnormal units. As a result, the MDBUTMF operation has the outstanding performance and less complexity for Consistent Sizableness Impulsive Outlier (CSIO) however the limitation of the original MDBUTMF is the recreation of the outlier units where all illustrational units in the window are outlier and the recreational unit is computed from all outlier illustrational units therefore the original MDBUTMF has the limited performance, especially at heavy outlier percentage. The algebraically article proposes the novel outlier extraordinariness sensing and recreation operation established on the adaptive MDBUTMF (Modified decision based unsymmetric trimmed median filter) which is promoted from the original MDBUTMF. In the computer simulation, numberless outlier illustrations, which are coalesced of GIRL and BABOON at all outlier percentage, are considered by the proposed adaptive MDBUTMF operation in contradistinction of the performance of state-of-art schemes: SMF, MF, AMF, WMF and original MDBUTMF. Due to the simulation results, the performance of the proposed adaptive MDBUTMF operation is exceeding than all preceding operations at all outlier percentage in both PSNR and illustrational quality, especially at heavy outlier percentage.
Authors
- Vorapoj Patanavijit (ORCID: https://orcid.org/0000-0002-9122-1276)
- Darun Kesrarat (ORCID: https://orcid.org/0000-0002-9931-1073)
- Kornkamol Thakulsukanant (ORCID: https://orcid.org/0000-0002-8196-1537)
- Wilaiporn Lee (ORCID: https://orcid.org/0000-0001-7484-0246)
- Kanabadee Srisomboon (ORCID: https://orcid.org/0000-0002-3120-3324)
Institutions
- Assumption University (TH)
- King Mongkut's University of Technology North Bangkok (TH)
Publication Details
- Published
- 2025-03-05
- DOI
- https://doi.org/10.1109/ieecon64081.2025.10987597
- Primary Topic
- Anomaly Detection Techniques and Applications
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- Innovation Fund