Automatic Waste Classification System using Deep Leaning Techniques

Waste management refers to a system that starts with classifying different kinds of waste and gradually managing it from its inception to its final disposal. This type of study can bring about a positive change as it will help making the environment pollution free and reuse the waste as much as we can by classifying the recyclable stuff from the waste that are considered useless. This paper proposes a custom CNN model that classifies different types of waste materials accurately. Here, a large dataset named "Garbage Classification" was used and 8 different classes: battery, biological, cardboard, clothes, green-glass, paper, plastic, trash - have been detected. The images have been augmented in order to make all the classes equal in size which has resulted in a total of 16,000 images. Pre-trained CNN models such as VGGNet16, Resent50, MobileNetV2, InceptionV3 along with custom CNN models have been used and successfully achieved 87.57%, 94.34%, 96.99%, 95.71%,97.16% train accuracy and 89.38%, 94.34%, 96.81%, 94.47% and 97.58% validation accuracy respectively. Later on, the paper also evaluates the custom CNN model’s performance on an unseen test dataset via confusion matrix.