Radix bupleuri seed identification method based on attention mechanism

Abstract

The invention discloses a radix bupleuri seed identification method based on an attention mechanism, which combines the attention mechanism, fuses four layers of feature maps through up-sampling, outputs a semantic flow in the fusion process, and applies a deep learning framework with a plurality of attention mechanisms to different types of radix bupleuri seed classification tasks for the first time. Good accuracy is realized, the deep neural network based on the attention mechanism ensures that no degeneration phenomenon is generated during training, and the neural network can adaptively adjust the weights of the semantic flow and the spatial flow due to the introduction of the attention mechanism. Mutual fusion of global features and local detail texture information during detection is ensured, so that a neural network generates a good classification effect.

Description

Technical field [0001] The present invention belongs to the field of medical image processing, involving a method of calculating the firewood seed recognition based on the attention mechanism. Background technique [0002] Chinese medicine is a special drug unique to traditional Chinese medicine, and its quality is critical to the efficacy of Chinese medicine prescriptions. It is essential to plant traditional Chinese medicines with quality, high quality requirements for traditional Chinese medicine seeds. At present, my country's traditional Chinese medicine planting base has generally not formulated standardized management, selection, and seedling without uniform industry standards. Artificial screening quality seed seedlings need to spend a lot of human cost, screening the subjectivity, difficult to solve the problem of uneven quality of traditional Chinese medicine. To this end, modern information technology promotes the standardization of traditional Chinese medicine plantin...

Images