RealWaste Image Multiclass Classification

Overview

This project focuses on building a deep learning model for multiclass image classification of waste materials using the RealWaste dataset, which was sourced from waste received at the Whyte's Gully Waste and Resource Recovery facility in Wollongong, NSW, Australia. The project explores both Deep CNNs and Transfer Learning techniques. Models are trained on the training dataset and evaluated on the validation set. The best-performing model, DenseNet121-FT, is further tested on the test set. Transfer learning is implemented using two approaches: feature extraction and fine-tuning to optimize performance.

---

Dataset

The dataset comprises 4,752 labeled images of waste, with each image sized at 524x524 pixels. The dataset is publicly available from the UC Irvine Machine Learning Repository. It is a comprehensive dataset that covers various classes of landfilled waste for sustainable waste management:

Label	Image Count
Cardboard	461
Food Organics	411
Glass	420
Metal	790
Miscellaneous Trash	495
Paper	500
Plastic	921
Textile Trash	318
Vegetation	436

• Publication Reference:
  Single, S.; Iranmanesh, S.; Raad, R. RealWaste: A Novel Real-Life Data Set for Landfill Waste Classification Using Deep Learning. Information 2023, 14, 633.
  DOI: 10.3390/info14120633

• Dataset Source:
  UC Irvine Machine Learning Repository

  Dataset URL: https://archive.ics.uci.edu/static/public/908/realwaste.zip

---

Project Highlights

1. Dataset Exploration - Tackling dataset imbalanceness

The RealWaste dataset is explored ,providing insights into the distribution and characteristics of the data. Given the imbalanceness of the dataset the class_weight is implemented to force the model to pay more/less attention to the classes based on their frequency.

2. Data Augmentation

Data augmentation is a technique used to artificially increase the size of a training dataset by applying various transformations to the existing data. It usually adopted to ensure the robustness and generalization of the model.

In this specific case I will apply the following transformations: * Horizontal flip * Rotate * Shear

3. Model Development

   Different type of models are implemented, trained and evaluated.

   • Convolutional Neural Networks:

     • Base-DeepCNN: A simple architecture with three feature extraction blocks (convolution + pooling + batch normalization) followed by a classifier.
     • Deep-CNN with Residual Connections and Depthwise Separable Convolutions: Incorporates residual connections to mitigate vanishing gradient issues and depthwise separable convolutions for a smaller, faster-converging, and less overfitting-prone model.

   • Transfer learning: Transfer learning consists in reusing a model which has been already trained on a large dataset (and thus it is a generic model) for our purposes. I adopted the two pre-trained models that performed the best in the original paper:

     • InceptionV3
     • DenseNet121

---