Identifying Fagaceae and Lauraceae species using leaf images and convolutional neural networks |
| |
| Institution: | 1. Department of Biomechatronics Engineering, National Taiwan University, Taipei, Taiwan;2. The Experimental Forest, College of Bio-Resources and Agriculture, National Taiwan University, Nantou, Taiwan;3. Department of Forestry, National Pingtung University of Science and Technology, Pingtung, Taiwan;4. School of Forestry and Resource Conservation, National Taiwan University, Taipei, Taiwan;1. CAS Key Laboratory of Tropical Forest Ecology, Xishuangbanna Tropical Botanical Garden, Chinese Academy of Sciences, Menglun 666303, China;2. Center for Plant Ecology, Core Botanical Gardens, Chinese Academy of Sciences, Xishuangbanna 666303, China;3. Global Change Research Group, Xishuangbanna Tropical Botanical Garden, Chinese Academy of Sciences, Menglun 666303, China;4. Department of Life Sciences, University of Chinese Academy of Sciences, Beijing 100049, China;1. Fisheries Resources Consultant, Food and Agriculture Organization of the United Nations, Viale delle Terme di Caracalla, 00153 Rome, Italy;2. Elasmobranch Research, Rehaegenstraat 4, 2820 Bonheiden, Belgium;3. Fin Forensics, 1145 10th Ave #322, Seattle, WA 98102, USA;4. Stick Figure Fish Illustration, 22 Lancewood Ave, Peregian Beach, Queensland 4573, Australia;5. Ludwig-Maximilians-Universität München, Department Biology II, Aquatic Ecology, Großhaderner Str. 2, 82152 Planegg-Martinsried, Munich, Germany;6. Department of Biology, University of Padova, Viale G. Colombo 3, 35131 Padova, Italy;7. Departamento de Ecología y Biología Animal, Universidad de Vigo, Campus Universitario de Vigo, 36310 Vigo (Pontevedra), Spain;1. School of Environment and Resource, Southwest University of Science and Technology, Number 59, Middle of Qinglong Road, Fucheng District, Mianyang 621-010, Sichuan, China;2. Jinniu District Administrative Examination and Approval Bureau of Chengdu, Number 77, South of Jinke Second Road, Jinniu District, Chengdu 610-000, Sichuan, China;3. School of Life Science and Engineering, Southwest University of Science and Technology, Number 59, Middle of Qinglong Road, Fucheng District, Mianyang 621-010, Sichuan, China;1. Department of Statistical Sciences, University of Cape Town, 8001 Cape Town, South Africa;2. Department of Biological Sciences, University of Cape Town, 8001 Cape Town, South Africa;1. College of Computer & Information Engineering, Central South University of Forestry and Technology, 410004 Changsha, Hunan, China;2. National University of Defense Technology, 410015 Changsha, Hunan, China;3. HuangFengQiao State-Owned Forest Farm, YouXian County, 412300 Zhuzhou, Hunan, China;4. Plant Protection Research Institute, Academy of Agricultural Sciences, 410125 Changsha, Hunan, China;1. College of Engineering, Northeast Agricultural University, Harbin 150030, China;2. Heilongjiang Green Food Development Center, China;3. College of Arts and Sciences, Northeast Agricultural University, Harbin 150030, China |
| |
| Abstract: | Lauraceae and Fagaceae are two large woody plant families that are predominant in the low- and middle-altitude regions in Taiwan. The highly interspecific similarity between some species of the family brings limitations on the management and utilization. This work proposed an approach for identifying 15 Lauraceae species and 20 Fagaceae species using leaf images and convolutional neural networks (CNNs). Leaf specimens of 35 species were collected from the northern, central, and southern parts of Taiwan. Images of the leaves were acquired using flat-bed scanners. Three CNN architectures—DenseNet-121, MobileNet V2, and Xception—were trained. Xception achieved the highest mean test accuracy of 99.39%, and MobileNet V2 required the shortest mean test time of 17.1 ms per image using a GPU. The saliency maps revealed that the characteristics learned by models matched the leaf features used by botanists. A pruning algorithm, gate decorator, was applied to the trained models for reducing the number of parameters and number of floating-point operations of the MobileNet V2 by 55.4% and 69.1%, respectively, while the model accuracy was maintained at 92.03%. Thus, MobileNet V2 has the potential to be used for identifying the Lauraceae and Fagaceae species on mobile devices. |
| |
| Keywords: | |
| 本文献已被 ScienceDirect 等数据库收录！ |
|
