Unlock instant, AI-driven research and patent intelligence for your innovation.

A method and device for automatic statistical analysis of test paper scores based on m-cnn

A statistical analysis and scoring technology, applied in the field of image processing, can solve the problems of fixed image size, waste of computing resources, and test paper scores cannot meet the requirements, so as to reduce the computational burden and achieve the effect of fast and accurate identification.

Active Publication Date: 2021-10-12
HENAN POLYTECHNIC UNIV
View PDF7 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, the LeNet-5 model also has the limitations of convolutional neural networks: the input image size is required to be fixed, and only a single handwritten digit can be recognized
Directly applying the LeNet-5 model to realize the statistics of test paper scores obviously cannot meet the requirements. At the same time, when the convolution layer in the model is operated, it is meaningful to convolute the area with non-zero pixel values ​​in the handwritten digit area. The convolution of the zero area is obviously a waste of computing resources

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • A method and device for automatic statistical analysis of test paper scores based on m-cnn
  • A method and device for automatic statistical analysis of test paper scores based on m-cnn
  • A method and device for automatic statistical analysis of test paper scores based on m-cnn

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0074] In order to solve the problem of efficient automatic statistics of test paper scores, the present invention provides an automatic test paper score statistics method based on M-CNN (Mask-CNN), which performs convolution, pooling, and other operations on the digital area according to Mask, and the final The pooling results are input into the SPP network, and finally the fast and accurate recognition of numbers is realized, and the final recognition results are counted. The overall process of the system is as figure 1 shown, including the following steps:

[0075] 110. Create a single-digit handwritten digit database based on the Mnist handwritten digit database, and use the NIST SD19 data set to create a double-digit database; the single-digit image of each single-digit handwritten digit database is called a single-digit handwritten sample, and the double-digit images in each double-digit database The digital images are called dual-digit handwriting samples.

[0076] Th...

Embodiment 2

[0100] A kind of test paper score automatic statistical analysis device based on M-CNN, it is the virtual device of embodiment, please refer to Figure 16 shown, which includes:

[0101] The first creation module 210 is used to create a single-digit handwritten digital library according to the Mnist handwritten digital library, and applies the NISTSD19 data set to create a double-digit library; the single-digit image of each single-digit handwritten digital library is called a single-digit handwritten sample, and each double The double-digit images in the digital library are called double-digit handwriting samples;

[0102] The extraction module 220 is used to extract non-zero pixels in each single-digit handwriting sample and double-digit handwriting samples, the non-zero pixels of each single-digit handwriting sample form a single-digit Mask area, and the non-zero pixels of each double-digit handwriting sample Form a double-digit Mask area, and obtain the pixel position of ...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses an M-CNN-based automatic statistical analysis method for test paper scores, which comprises the following steps: Step 1, creating a single-digit handwritten digital database and a double-digit database; Step 2, obtaining each single-digit Mask area and double-digit The pixel position of the Mask area; Step 3, create a single-digit CNN recognition model and a double-digit CNN recognition model; Step 4, train the single-digit CNN recognition model and a double-digit CNN recognition model; Step 5, obtain a target image; Step 6. Cut the target image to determine the image to be recognized; Step 7. Acquire the value of each image to be recognized; Step 8. Obtain the total score of the test paper image. The present invention also provides an M-CNN-based automatic statistical analysis device for test paper scores. The invention can effectively reduce part of the calculation burden, and at the same time introduce the SPP network to effectively solve the problem of fixed size of the input image.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to an M-CNN-based automatic statistical analysis method and device for examination paper scores. Background technique [0002] Examination is an effective evaluation method of teaching effect and an effective reference method for the improvement of teaching activities. Therefore, the statistical analysis of the scores in the test questions is particularly important. With the continuous emergence of handwritten digit recognition algorithms based on convolutional neural networks and the continuous improvement of recognition accuracy, the application of handwritten digits based on CNN (Convolutional Neural Network) is becoming more and more extensive. As a classic algorithm for handwritten digit recognition, LeNet-5 has high recognition accuracy and simple model, which has certain advantages in practical applications. However, the LeNet-5 model also has the limitations of co...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): G06K9/62G06K9/68
CPCG06V30/2455G06F18/23
Inventor 赵运基陈相均张新良王加朋张海波范存良
Owner HENAN POLYTECHNIC UNIV