当前位置: X-MOL 学术Neural Process Lett. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Deep Learning Neural Network for Unconventional Images Classification
Neural Processing Letters ( IF 2.6 ) Pub Date : 2020-04-23 , DOI: 10.1007/s11063-020-10238-3
Wei Xu , Hamid Parvin , Hadi Izadparast

The pornographic materials including videos and images are easily in reach for everyone, including under-age youths, allover Internet. It is also an aim for popular social network applications to contain no public pornographic materials. However, their frequent existence throughout all the Internet and huge amount of available images and videos there, make it impossible for manual monitoring to discriminate positive items (porn image or video) from benign images (non-porn image or video). Therefore, automatic detection techniques can be very useful here. But, the traditional machine learning models face many challenges. For example, they need to tune their many parameters, to select the suitable feature set, to select a suitable model. Therefore, this paper proposes an intelligent filtering system model based on a recent convolutional neural networks where it bypasses the aforementioned challenges. We show that the proposed model outperforms the recent machine learning based models. It also outperforms the state of the art deep learning based models.

中文翻译:

深度学习神经网络用于非常规图像分类

包括视频和图像在内的色情材料很容易在互联网上为每个人(包括未成年人)提供。这也是流行的社交网络应用程序不包含任何公共色情材料的目的。但是,它们在整个Internet上频繁存在,并且在那里存在大量可用的图像和视频,因此手动监视不可能将阳性项目(色情图像或视频)与良性图像(非色情图像或视频)区分开。因此,自动检测技术在这里可能非常有用。但是,传统的机器学习模型面临许多挑战。例如,他们需要调整其许多参数,选择合适的功能集,选择合适的模型。因此,本文提出了一种基于最近的卷积神经网络的智能过滤系统模型,该模型绕过了上述挑战。我们表明,提出的模型优于最近的基于机器学习的模型。它还优于基于深度学习的模型。
更新日期:2020-04-23
down
wechat
bug