Implementation of Faster Region-Based Convolutional Neural Network for Waste Type Classification

The improper disposal of garbage is one of the factors that contribute to the problems that are visible in the environment, and these problems have a significant influence. The digestive issue is one of the problems caused by the increase in the population of insects and the variety of different types of pests. If waste is not recycled and managed effectively, it can negatively impact economic growth and contribute to air pollution and other health issues. Integrating computer technology, such as object recognition, might be helpful in waste management. At this point, the most popular approach to object detection is called deep learning. Therefore, this paper presents a waste recognition using Faster Region-Based Convolutional Neural Network. The research focused on recognizing plastic and disposable diapers as two separate classes since disposable diapers and plastics are essential in modern society. Convolutional neural network (CNN) which comes under deep learning has been widely used in object recognition or classification due to its effectiveness. In the Faster RCNN, we used two separate networks to predict the region proposal reshaped by a Region of Interest (RoI) pooling layer to classify the images. From the result, it can be observed that our proposed system has achieved an average accuracy of 91.2%.