Explored advanced machine learning techniques to optimize CNNs for waste classification, employing ablation studies on data augmentation, model architecture, and dataset size to enhance practical deployment considerations.
Successfully deployed our classification model on a Raspberry Pi 5, demonstrating the application of sophisticated machine learning strategies in solving real-world engineering problems within an environmental sustainability context.
Embedded Files
After ablation studies, the results of which dictated training transforms and considerations, the results on three selected architectures can be seen above.
MLAI Final Report.pdfPage updated
Google Sites
Report abuse