Digital Repository

Harmful Algae Image Classification With Machine Learning

Show simple item record

dc.contributor.author Ali, Abdikani
dc.date.accessioned 2021-09-01T18:23:32Z
dc.date.available 2021-09-01T18:23:32Z
dc.date.created 2021
dc.identifier.uri http://hdl.handle.net/11216/3958
dc.description 2021 Celebration of Student Research and Creativity presentation en_US
dc.description.abstract "Harmful algae can grow rapidly into blooms and release toxins that can contaminate drinking water sources which can be unsafe to humans and wildlife. This research has discovered images belonging to seven different genera of harmful algae which were used as a data set to train a convolutional neural network (CNN) to automatically classify harmful algae at the microscopic level. After data augmentation, 20,010 images were trained on the network. Our CNN has shown an overall accuracy of 99.75%. The next steps include collecting more original images and modification the neural network architecture to improve the accuracy of unseen images." en_US
dc.description.uri https://youtu.be/_ITLOd7DLFg en_US
dc.language.iso en_US en_US
dc.publisher Northern Kentucky University en_US
dc.relation.ispartofseries Celebration of Student Research and Creativity;2021
dc.subject Toxic algae Toxicology en_US
dc.subject Algal toxins en_US
dc.subject Machine learning en_US
dc.title Harmful Algae Image Classification With Machine Learning en_US
dc.type Presentation en_US


Files in this item

Files Size Format View

There are no files associated with this item.

This item appears in the following Collection(s)

Show simple item record