论文信息 - Mobile Leaf Identification System using CNN applied to plants in Hokkaido

Mobile Leaf Identification System using CNN applied to plants in Hokkaido

The ability to identify plant types is important when conducting vegetation surveys. This ability requires investigators experience. We propose a mobile application using convolutional neural networks (CNNs) that will help beginners identify plant species. We compare three CNN models, VGG19, MobileNet, and MobileNetV2. Our plant identification application using MobileNetV2 shows an average F1 score of 0.992, indicating its high performance and practicality. The implemented system shows a practical performance of 338.1 ms per picture on a tablet-type device.

[1] Mark Sandler,et al. MobileNetV2: Inverted Residuals and Linear Bottlenecks , 2018, 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.

[2] Andrew Zisserman,et al. Very Deep Convolutional Networks for Large-Scale Image Recognition , 2014, ICLR.

[3] Yucel Inan,et al. Leaves Recognition System Using a Neural Network , 2016 .

[4] Sang-Yong Rhee,et al. Plant Leaf Recognition Using a Convolution Neural Network , 2017, Int. J. Fuzzy Log. Intell. Syst..

[5] Bo Chen,et al. MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications , 2017, ArXiv.

[6] Abhishek Das,et al. Grad-CAM: Why did you say that? , 2016, ArXiv.