Theses & Dissertations

Permanent URI for this collection

https://hdl.handle.net/20.500.14170/209

Browse

Browsing Theses & Dissertations by Author "Afiq Ahmad Shakri"
  • Publication
    Biological viruses images recognition using artificial intelligence classifier
    ( 2018)
    Afiq Ahmad Shakri
    This thesis uses Artificial Intelligence for biological viruses images classification. The start of the development of this system requires a number of biological viruses images which belongs to different classes for this research work. The first stage of this system before the virus can be classified are as such, the images first needs to be preprocessed using noise density of different level. It is crucial because in biological viruses images classification, the detail of the viruses images are important due to how closely some of the viruses resembles each other even though they are not in the same classes type of virus. Therefore, image preprocessing is required for ensuring the feature extraction would be able to correctly extract the viruses details allowing the classification algorithm to produce a high accuracy results. This process are done by smearing the images with noise. For this research work, salt&pepper noise are used in the system development. The second stages are after the noise have been applied to the images, several feature extraction methods have been developed to extract the images feature out. Feature extraction method plays an important part in this research for determining and acquiring the best possible results accuracy. The third stage of this experiments required several artificial intelligence classification algorithms. The algorithm will classify and produce the results accuracy based on the extracted feature from feature extraction method. This will show the accuracy of the feature extraction method is based on how accurate the biological viruses are classified. The accuracy is based on the combination of feature extraction method and artificial intelligence classification algorithm. The real time experiment conducted proved that the proposed feature and classifier combination are robust, excellent, and efficient of which it has produced a results accuracy of up to 99.93% for biological viruses images classification. The proposed combination produced a much better result as compared with most of the real time applications of this system.