Teaching Figure Collection based on Baidu API and Deep Learning Models

This paper introduces a deep learning-based teaching figure collection method that cooperates with Baidu API’s general object recognition service. The Baidu API is to output the general information of the figures. Then, we use deep learning models to classify the specific information of figures. Finally, we fuse these to correctly classify the figures, so that, we can store these figures to the correct directory. We utilized some public datasets for the experiment, which include CIFAR-10 dataset, CIFAR-100 dataset, and Mini-ImageNet dataset. Furthermore, we apply our method to the collection of teaching figures to prove the scalability.