Under the joint guidance of Associate Professor Song Haiyu and Dr. Wu Haiyan, Wang Haoyu, a ZUFE undergraduate student of Computer Science Class 1, Grade 2019, School of Information Management and Artificial Intelligence, published a research paper entitled “ Multilayer Feature Fusion and Attention-based Network for the JCR Region II Journal, Journal of Plant Diseases and Protection Crops and Weeds Segmentation. .

Journal of Plant Diseases and Protection is an international scientific journal in the field of comprehensive application in agriculture and forestry. It mainly publishes articles on applied sciences such as plant pathology, plant health, and plant protection, especially the application research results of advanced technologies such as artificial intelligence in interdisciplinary fields of agriculture and forestry.

Distinguishing weeds from crops is a critical challenge in agriculture, with the existing agriculture semantic segmentation networks simply combining low-level with high-level features at the encoder and decoder stages to improve performance. This paper proposes a novel dual attention network (DA-Net), based on branch attention blocks in the encoding stage and spatial attention blocks in the decoding stage, to bridge the gap between low-level and high-level features. Then, a cascaded convolution block utilizing asymmetric convolution is constructed, supporting the receptive field’s expansion without increasing the computational burden or the parameter cardinality. It designs a spatial attention block in the fusion stage to capture rich contextual dependencies. Finally, a novel block named densely channel fusion is constructed, which utilizes a sub-pixel layer to encode most channel information into spatial information. The experimental results demonstrate that DA-Net is superior to ExFuse, Ddeeplabv3(+), and PSPNet on three public datasets, with each added component significantly affecting the overall performance.

Since Wang Haoyu joined Song Haiyu’s research team in his freshman year, he has studied diligently and actively explored the interdisciplinary research of “artificial Intelligence +”. He has won more than 10 international and national awards in various discipline competitions and published a JCR II paper. He has submitted another paper to an SCI II journal (as the lead author, under review), and applied for two national invention patents.