Enhancing Waste Classification with Convolutional Neural Networks and Explainability Techniques: Power of Computer in Classification
Kakul Gupta, Pragya Arora, Raman Mohan Sharma, Noor Yeshfeen, Ritika Kumari, Poonam Bansal · 2023
India is having a huge population and due to this it suffers from a lot of human wastage disposal problem to combat this issue a good waste Segregation model is required. Currently there is no system which can divide the waste into more than two categories in real life scenario for Indian population to solve this in our study we are dividing our waste label into eight categories These categories are according to the industrial requirements and now as the model should be reliable, we are using explicability AI models like LIME SHAP AND GRAD CAM. This would help to make the model accountable and reliable so that the big organization can also use this and it would be also suitable for the waste Segregation of daily household waste materials. In this study using convolution neural networks, classification of different waste materials are done, as it is a deep neural network it is giving a good accuracy.