Research And Implementation Of Laptop BIOS Automated Test Based On Deep Learning

With the development of computer technology,the functions of laptop have become more powerful.But the composition structure has become more and more cumbersome,which brings challenges to its stability test.As the bottom layer of the computer,BIOS(Basic Input Output System)is directly connected to the main hardware devices.The traditional method of manually testing BIOS has gradually been unable to meet the current needs for laptop stability testing.Therefore,the application of automated testing can effectively achieve an efficient and reliable testing process.Since the existing automated test tools cannot be run directly in the BIOS,this paper uses deep learning technology to analyze the image of the BIOS interface.Realize the automatic control through the BIOS interface text recognition,text detection and laptop boot logo recognition.The main work and results are as follows:(1)In order to recognize the text information in the BIOS interface,this paper discusses the structure of the text recognition model CRNN,and optimizes the CRNN structure with the goal of reducing the amount of model parameters and improving the recognition accuracy.A text recognition network SSS-Net based on convolutional neural network and attention mechanism is designed.It integrates the strip convolution and attention mechanism,which are added to the text recognition framework in a modular manner.The network includes fringe convolution and attention mechanisms,and they are assembled into a network model in the form of functional modules.The performance of horizontal text recognition is verified in the public data sets.Compared with the CRNN framework with the same convolutional neural network feature extraction part,SSS-Net has higher recognition accuracy,and the parameter amount is only 1/4 of CRNN.(2)In order to achieve the functional requirements of automatically entering the BIOS,a Logo recognition network based on Res Net is designed,and the SE module is added to the residual block structure.According to the performance of the constructed BIOS data set,the specific network framework structure is determined.(3)In order to locate the specific position of the text line in BIOS,a text detection framework based on convolutional neural network and recurrent neural network is constructed.Among them,VGG16 is used for feature extraction,and Bi-directional Gated Recurrent Unit(Bi-GRU)is used for sequence information integration.The text detection framework and the text recognition framework SSS-Net are spliced together to form a complete OCR(Optical Character Recognition)framework to complete the task of text detection and recognition on the BIOS interface.According to the experimental results in the constructed BIOS data set and the test performance applied in practice,the actual application performance of the OCR framework is verified.(4)Combine the Logo recognition framework and OCR framework,and cooperate with other hardware devices to build the Biz Bios Test for BIOS automated testing.And realize the fully automated test process of BIOS.This article first compares the performance of the text recognition network SSS-Net and CRNN in the public data set.Experimental results show that the text recognition framework proposed in this paper is more sensitive to the specific position of characters than CRNN,and effectively acquires image text features with lower parameters.Exploring a new recognition method for image text.Then,the text recognition SSS-Net,text detection framework and laptop logo recognition framework constructed in this paper are trained in the constructed BIOS data set and applied to practical tasks.The experiment verifies their performance on BIOS interface text detection and text recognition tasks and Logo recognition tasks.Finally,with the hardware equipment,this paper realizes the fully automatic control of the laptop BIOS,which greatly saves the labor cost and improves the test efficiency.