Amodal segmentation of cane sugar crystal via deep neural networks

The growth of cane sugar crystals can provide fundamental information for the setting of production process parameters. However, the overlap phenomenon of sugar particles has adverse effects on identifying the shape of sugarcane crystals in images, and it is critical to effectively recognize the status of the growth of cane sugar crystals. To realize the recognition and segmentation of three cane sugar crystal states, we propose ASSugarNet, a cane sugar crystal amodal segmentation method based on deep neural networks. To better extract multi-scale semantic features, the local feature enhance module is proposed. The foreground boundary enhancement loss function is adopted, which can effectively improve the segmentation quality of sugar crystal boundary regions. For overlapping crystals, the amodal segmentation module is used to predict the contours and masks of the occluded sugar crystals, which provides accurate input for the subsequent sugar crystal growth stage judgment. To verify the network performance, we create a dataset of cane sugar crystals. Compared with FCN, PSPNet and U-Net 3 + network, the experimental results show that the proposed network structure has better segmentation for three types of crystal performance, with MIoU of 81.47% and OA of 91.00%. The AP of ASSugarNet for adhesive particle segmentation is 32.25%, and the AP50 is 52.14%.

Automated summary

The growth of cane sugar crystals is crucial for the production process. However, the overlap of sugar particles hampers the accurate identification of crystal shapes from images. To address this, we propose ASSugarNet, a deep neural network-based amodal segmentation method for cane sugar crystals. The network incorporates a local feature enhance module and a foreground boundary enhancement loss function to improve segmentation quality. It also uses an amodal segmentation module to predict contours and masks for occluded crystals. Experimental results demonstrate that ASSugarNet outperforms other networks in crystal segmentation, achieving an MIoU of 81.47% and an OA of 91.00%. The AP for adhesive particle segmentation is 32.25%, with an AP50 of 52.14%.