What question did this study set out to answer?

The aim is to improve waste classification accuracy using deep learning models on image data.

April 3, 2026MathematicsOpen Access

Image-Based Waste Classification Using a Hybrid Deep Learning Architecture with Transfer Learning and Edge AI Deployment

View Full Paper

Ask AI

Bookmark

Authors

DVDomen VerberUniversity of Maribor TGTeodora GrnevaUniversity of Maribor JDJani DugonikUniversity of Maribor

Discussion

Member takes

Implication

This research demonstrates enhanced waste classification accuracy using a hybrid model on image data, indicating practical AI deployment capabilities.

Key Points

The aim is to improve waste classification accuracy using deep learning models on image data.
Investigated three CNN architectures (ResNet-50, EfficientNet-B0, MobileNet V3) and a custom hybrid model (CustomNet).
Utilized a dataset of 13,933 RGB images categorized into 10 waste classes.
Applied transfer learning and preprocessing techniques like normalization and data augmentation.
Evaluated performance using metrics such as accuracy, macro-F1, and ROC–AUC.
CustomNet achieved 97.79% accuracy with a macro-F1 score of 0.973 and ROC–AUC of 0.992.
CustomNet outperformed other models (p<0.001) and matched ResNet-50 performance at a lower parameter count.
Practical deployment on Edge AI platforms was feasible with real-time inference speed of 86.70 ms per image.

Cite This Study

Verber et al. (2026) studied this question.

synapsesocial.com/papers/69cf5fe05a333a821460eaa9 https://doi.org/10.3390/math14071176

View Full Paper

Ask AI

Bookmark

Implication

This research demonstrates enhanced waste classification accuracy using a hybrid model on image data, indicating practical AI deployment capabilities.

Image-Based Waste Classification Using a Hybrid Deep Learning Architecture with Transfer Learning and Edge AI Deployment

Discussion

Member takes

Implication

Key Points

Cite This Study

Image-Based Waste Classification Using a Hybrid Deep Learning Architecture with Transfer Learning and Edge AI Deployment

Discussion

Member takes

Implication

Key Points

Cite This Study

Also Consider

Also Consider