Waste Classification Dashboard Using TensorFlow and Convolutional Neural Network (CNN) for Sustainable Waste Management

Effective waste management represents a significant global challenge in the ongoing effort to preserve environmental cleanliness and public health. A critical component of this process is the accurate classification of waste, which directly facilitates more efficient recycling and disposal systems. This research focuses on the development and implementation of a sophisticated waste classification model utilizing a Convolutional Neural Network (CNN). The model is built using the TensorFlow framework and subsequently optimized with TensorFlow Lite (TFLite) for lightweight integration into a web-based dashboard powered by Flask. The dataset for this study comprises a diverse collection of waste images, which has been artificially expanded through data augmentation techniques to enhance the model's ability to generalize. The CNN architecture, specifically a ResNet50 model, was meticulously trained with multiple layers to effectively extract distinguishing features from the images, achieving an optimal classification accuracy of 91%. For practical deployment, the trained model was converted to the TFLite format, enabling efficient performance on devices with limited computational resources. The final implementation allows users to upload images of waste items and receive real-time classification results through the interactive web interface. The system demonstrated a success rate of 67.33% in detecting various types of waste during testing. This study successfully demonstrates the significant potential of deep learning methodologies to automate the waste identification process, thereby contributing to the creation of more intelligent and effective waste management systems that can be used

Keywords: CNN, TensorFlow, TFLite, Flask, Deep Learning, Waste Classification