Please use this identifier to cite or link to this item: https://rda.sliit.lk/handle/123456789/245
Full metadata record
DC FieldValueLanguage
dc.contributor.authorAhmed Akmal, M. A-
dc.date.accessioned2021-10-25T07:00:00Z-
dc.date.available2021-10-25T07:00:00Z-
dc.date.issued2021-05-
dc.identifier.urihttp://localhost:8080/jspui/handle/123456789/245-
dc.description.abstractBiomedical wastes (BMWs) include potentially infectious, sharps, pharmaceuticals and radioactive wastes probably generated by hospitals, vaccination centers, biomedical laboratories, etc. Handling and disposal of biomedical wastes potentially have multiple risk factors. Currently, hospitals and laboratories use color-coded bins to classify and categorize different types of wastes to ease the handling and the disposal process. Sometimes due to human errors these wastes could be miscategorized or misplaced in different bins. In recycling terms this is known as waste contamination. Contaminating the biomedical waste streams causes a huge potential threat to the people who handle them. Computer vision based biomedical waste classification is one of the best ways to prevent these issues. But applying pure computer vision algorithms is much more suitable for small tasks such as pattern recognition, edge detection etc. In order to classify different kinds of biomedical wastes, then convolutional neural networks (CNN) would be a much more suitable choice. This research proposes a deep learning model which accurately classifies several selected biomedical wastes such as syringes, blades and sample collection tubes with a prediction accuracy around 96% on the test dataset. Further the implemented model approximately localizes the biomedical wastes to serve robotics and smart-bin applications.en_US
dc.language.isoenen_US
dc.subjectBiomedicalen_US
dc.subjectWaste Sortingen_US
dc.subjectClassificationen_US
dc.subjectUsing Deep Learningen_US
dc.titleBiomedical Waste Sorting & Classification Using Deep Learningen_US
dc.typeThesisen_US
Appears in Collections:2021
MSC IT

Files in This Item:
File Description SizeFormat 
MS19806778.pdf
  Until 2050-12-31
9.19 MBAdobe PDFView/Open Request a copy
MS19806778-pages-1-12.pdf400.83 kBAdobe PDFView/Open


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.