LS-Net: a convolutional neural network for leaf segmentation of rosette plants

Leaf segmentation from plant images is a challenging task, especially when multiple leaves are overlapping in images with a complex background. Recently, deep learning-based methods have demonstrated their effectiveness in the realm of image segmentation. In this study, a novel convolutional neural network called LS-Net has been proposed for the leaf segmentation of rosette plants. The experiment is performed over 2010 images from the plant phenotyping (CVPPP) and KOMATSUNA datasets. The segmentation ability of the LS-Net has been investigated by comparing it with four recently applied existing CNN-based segmentation models, namely DeepLab V3 + , Seg Net, Fast-FCN with Pyramid Pooling Module, and U-Net. The analysis of the experimental results clearly demonstrates the superiority of the proposed LS-Net to other tested CNN models.