A Simple Convolutional Neural Network for Doodle Classification

Visuals have been an integral part of our lives. The use of hand-drawn doodles is widespread and persistent, enabling people to retain information better and freely communicate thoughts and information. Doodle classification / hand-drawn diagram classification has significant application to communicate with persons who are deaf or hard of hearing that trains them to identify their urgent needs. The objective of this work is to develop a convolutional neural network (CNN) to classify the hand drawn images. This work utilized the Quickdraw dataset to train and evaluate the model. The performance of the customized model is compared with various state-of-the-art deep learning algorithms like MobileNet and DenseNet architectures and proved that proposed model comparatively gives better results on Quickdraw dataset.