Erotic Image Recognition Method of Bagging Integrated Convolutional Neural Network
Lizhi Huang, Xunyi Ren · Proceedings of the 2nd International Conference on Computer Science and Application Engineering · 2018
With1 the development of current Internet era, the exchange of data information is becoming more frequent and the spread of erotic images is becoming easier. Under these circumstances, it becomes even more important to identify and classify the images. Deep learning has been widely used in the field of image recognition because of its great advantage in automatically extracting features. However, in the case of small amount of data, it is easy to cause over-fitting of training data. Based on current situation, we propose a new erotic image recognition model. This model adopts Bagging Integrated Convolutional Neural Network and combines traditional Color Features-Histogram of Color based on the depth features. While improving the recognition accuracy, it also increases the sensitivity of the model to the color of the picture. Result of the experiment shows that, when identifying and classifying images in the NPDI data sets, the accuracy of the proposed model reaches 99.31%, which is 2.67% higher than that of the Convolutional Neural Network model, and it has a favorable classification recognition effect.