Fruit Classification based on ResNet and Attention Mechanism

Miaorun Lin

doi:10.54097/hset.v34i.5441

Authors

Miaorun Lin

DOI:

https://doi.org/10.54097/hset.v34i.5441

Keywords:

Fruit Classification, ResNet, Visual Attention, Residual Learning.

Abstract

Numerous types of fruits are discovered, which can provide humans with nutrients and trace elements. There are 59 families and 694 species of fruits in China, of which more than 300 kinds of cultivated fruit trees are in abundance, including more than 10000 varieties. Therefore, the classification of fruit detection is very necessary. The traditional method of fruit classification mainly relies on manual methods, which has greatly reduced the cost effectiveness in recent years due to the increase of labor cost. This paper introduces the idea of transfer learning by comparing various automated methods for detecting and classifying fruits, and incorporates the attention mechanism module into the ResNet model to achieve recognition of image features that can effectively classify fruits. This paper seeks to provide a complete review on fruit classification concluding that the ResNet model with the introduction of the attention mechanism has good accuracy and verifying the efficiency of the model in the field of image recognition and classification.

Downloads

Download data is not yet available.

References

He Zichao. Multi-task image classification based on deep learning and migration learning [D]. Guangzhou:South China University of Technology, 20177

Gomes et al., “Applications of computer vision techniques in the agriculture and food industry: a review”, Eur. Food Res. Technol., 235 (6), 989–1000. 2012.

Zhang et al., "Application of computer vision technology in agricultural field", Applied Mechanics and Materials, vol. 462. Trans Tech Publ, pp. 72–76, 2014.

Vibhute et al., "Applications of image processing in agriculture: a survey", Int. J. Comput. Appl., 52 (2), 34– 40, 2012.

Guo Hua. Deep learning and its significance [J]. Curriculum. Teaching materials. Teaching Methodology,2016(11):25-32.

https://www.kaggle.com/datasets/moltean/fruits

He K,Zhang XRen Set alDeep residual learningfor imagerecognition[C] Proceedings of the IEEE Conference on Computer Vision and PatternRecognition2016:770-778 Amjady N. Short-term hourly load forecasting using time series modeling with peak load estimation capability [J]. IEEE Transactions on Power Systems, 2001, 16(4): 798-805.

Guo ZY,Shu X,Liu CHANGYAN,et al. Flower recognition algorithm based on convolutional neural network with ReLU function[J]. Computer Technology and Development,2018(5):154-157,163.

Krizhevsky A, Sutskever I, Hinton G E. ImageNet classification with deep convolutional neural networks[J]. Communications of the ACM, 2017, 60(6): 84-90.

Wang WG, Shen JB, Jia YD. A review of visual attention detection[J]. Journal of Software,2019(2):416-439.