Deep Learning on Mobile Systems
暂无分享,去创建一个
Training of deep learning methods which are widely utilized in intelligent systems requires powerful computers, however, they are also used on mobile systems. Hence, application programming interfaces providing execution of pre-trained models on mobile systems have been introduced. In this study, two interfaces running on Android systems are tested on 40 objects for both object detection and labelling. Morever, ease of installations, resource usages were also compared. According to the test results, labelling performance of the model provided by Tensorflow interface is higher than the model of MLKit.
[1] Dumitru Erhan,et al. Going deeper with convolutions , 2014, 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).
[2] Sergey Ioffe,et al. Rethinking the Inception Architecture for Computer Vision , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).