Detection and Recognition Algorithm for Cigarette Box Documents based on Deep Learning
DOI:
https://doi.org/10.54097/fcis.v5i1.11544Keywords:
Text Detection, Text Recognition, Deformable Convolution, Deep LearningAbstract
In order to solve the problem of unclear detection and recognition caused by damage and distortion of the markings on the cigarette box during transportation, this article introduces an efficient channel attention mechanism SENet module for the DBNet detection model, effectively enhancing the model's detection ability. The introduction of deformable convolutional DCN on the CRNN recognition model further enhances the model's ability to detect distorted text. The experimental results show that the accuracy of the detection model reaches 88.43%, and the recall rate reaches 83. 47%, the comprehensive indicator reached 85.87%, with a significant improvement effect. The entire line recognition rate of the recognition model also reached 90.23%, which is better than the original algorithm.
Downloads
References
Tok W H ,Stéphane Bressan. DBNet: A Service-Oriented Database Architecture[C]//17th International Workshop on Database and Expert Systems Applications (DEXA 2006), 4-8 September 2006, Krakow, Poland. IEEE Computer Society, 2006. DOI:10.1109/DEXA.2006.48.
Schimanski L M , Drakesmith H , Merryweather-Clarke A T ,et al. In vitro functional analysis of human ferroportin (FPN) and hemochromatosis-associated FPN mutations[J].Blood, 2005, 105(10):4096-102.DOI:10.1182/blood-2004-11-4502.
Hu J, Shen L , Sun G .Squeeze-and-Excitation Networks [C]// 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).IEEE, 2018.DOI: 10.1109/ CVPR. 2018. 00745.
Zhang C , Woodland P .Parameterised Sigmoid and ReLU Hidden Activation Functions for DNN Acoustic Modelling [C]// Interspeech 2015.2015.DOI:10.21437/Interspeech.2015-649.
Mogren O .C-RNN-GAN: Continuous recurrent neural networks with adversarial training[J]. 2016.DOI: 10. 48550/ arXiv. 1611.09904.
Sundermeyer M ,Ralf Schlüter, Ney H .LSTM Neural Networks for Language Modeling[C]// Interspeech. 2012. DOI: 10. 1016/0165-6074(89)90269-X.
Graves A ,Santiago Fernández, Gomez F .Connectionist temporal classification: labelling unsegmented sequence data with recurrent neural networks[J].ACM, 2006.DOI: 10. 1145/ 1143844.1143891.
Tomáš Jeníček. Deformable Convolutional Networks [J]. [2023-08-19].
