- 期刊
MobileNet and EfficientNet Demonstration on Google Landmark Recognition Dataset
摘要
In image processing, the convolutional neural net- work extracts better features than the previous manual features because of the special structure of the CNN. The combination of convolution and pooling layers enables the CNN to extract better features in the image.There are many network models of convolution neural networks, but a convolution neural network model generally consists of several convolution layers, pooling layers and full connection layers. In this paper, we will discuss some newest models, MobileNet and EfficientNet, using Google Landmark Recognition's dataset to visualize these two CNN models' performance on image classification by comparing their accuracy and public score.
參考文獻
https://github.com/google/youtube-8m/blob/master/average precision calculator.py
https://www.kaggle.com/c/landmark-recognition-2020..
Howard, Andrew Zhu, Menglong Chen, Bo Kalenichenko, Dmitry Wang, Weijun Weyand, Tobias Andreetto, Marco Adam, Hartwig. (2017). MobileNets: Efficient Convolutional Neural Networks for Mo- bile Vision Applications.
Sinha, Debjyoti El-Sharkawy, Mohamed. (2019). Thin MobileNet: An Enhanced MobileNet Architecture.
Mingxing Tan, Quoc V. Le. (2019). EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks.