Muscarinic image recognition method and device based on deep residual network and transfer learning

Abstract

The invention discloses a muscarinic image recognition method and device based on a deep residual network and transfer learning. The method comprises the following steps: (1) collecting a muscarinic image, carrying out the foreground image extraction and data enhancement of the muscarinic image, carrying out the unified size processing, determining a classification label, and constructing a training set; (2) training a deep residual error network by utilizing the Image Net image set, and extracting deep residual error network parameters; (4) a muscarinic image recognition network is constructed, the muscarinic image recognition network comprises a convolution layer, a pooling layer, a full connection layer and a softmax classification layer, and deep residual network parameters are migrated to the convolution layer and the pooling layer; (5) training a muscarinic image recognition network by using the training set to obtain a trained muscarinic image recognition model; and (6) recognizing the to-be-recognized muscarinic image by using the trained muscarinic image recognition model to obtain a recognition result. The muscarinic image recognition method and device can accurately recognize and classify the muscarinic.

Description

technical field [0001] The invention belongs to the field of image recognition, in particular to a method and device for image recognition of toadstools based on deep residual network and transfer learning. Background technique [0002] Toadstools, also known as poisonous mushrooms and toadstools, refer to species that produce poisonous reactions to humans or livestock after the fruiting bodies of large fungi are eaten. There are 435 types of poisonous mushrooms in my country according to the literature. Because some of the poisonous mushrooms are very similar to edible wild mushrooms in shape, ordinary people do not have the professional ability to identify them, and they are easy to accidentally pick them up and eat them to cause poisoning. Therefore, how to accurately identify whether wild fungi are poisonous is a very critical issue for ordinary people and has important research significance. [0003] Mushroom identification methods mainly include morphological identifi...

AI Technical Summary

Problems solved by technology

These two methods aimed at the limitations of traditional toadstool identification, and established a toadstool identification model based on the structural attributes of toadstools. There are two shortcomings: one is due to factors such as the growth stage of toadstools and the environment in which they are located. It will lead to complex manual feature extraction, and the recognition results will depend on the quality of the feature selection of the mushrooms; on the other hand, the data sets of the mushrooms used in these two test methods come from the UCI database of the University of California, Irvine, and China The types of common toadstools in China are quite different, and the research on the identification of common toadstools in China is not applicable

Images