CN117632072A - Electronic paper display-oriented degraded document image mapping method - Google Patents
Electronic paper display-oriented degraded document image mapping method Download PDFInfo
- Publication number
- CN117632072A CN117632072A CN202311599318.1A CN202311599318A CN117632072A CN 117632072 A CN117632072 A CN 117632072A CN 202311599318 A CN202311599318 A CN 202311599318A CN 117632072 A CN117632072 A CN 117632072A
- Authority
- CN
- China
- Prior art keywords
- electronic paper
- image
- display
- mapping
- document image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000013507 mapping Methods 0.000 title claims abstract description 36
- 238000000034 method Methods 0.000 title claims abstract description 31
- 238000013135 deep learning Methods 0.000 claims abstract description 21
- 238000007781 pre-processing Methods 0.000 claims abstract description 13
- 230000009466 transformation Effects 0.000 claims abstract description 6
- 238000006243 chemical reaction Methods 0.000 claims description 14
- 230000009977 dual effect Effects 0.000 claims description 10
- CLSIFQGHPQDTHQ-DTWKUNHWSA-N (2s,3r)-2-[(4-carboxyphenyl)methyl]-3-hydroxybutanedioic acid Chemical compound OC(=O)[C@H](O)[C@@H](C(O)=O)CC1=CC=C(C(O)=O)C=C1 CLSIFQGHPQDTHQ-DTWKUNHWSA-N 0.000 claims description 9
- 230000000877 morphologic effect Effects 0.000 claims description 3
- 230000000694 effects Effects 0.000 abstract description 6
- 230000032683 aging Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 230000001502 supplementing effect Effects 0.000 description 1
- 230000002123 temporal effect Effects 0.000 description 1
Landscapes
- Control Of Indicators Other Than Cathode Ray Tubes (AREA)
Abstract
The invention provides a degraded document image mapping method for electronic paper display. The method comprises the following steps: firstly, an electronic paper display hardware platform is built, namely all devices are connected to enable the electronic paper to normally operate, and the electronic paper display hardware platform comprises a main control chip, an electronic paper driving board and an electronic paper display screen; secondly, an image processing system is built, and image preprocessing operation is carried out on the degraded document image, wherein the image preprocessing operation comprises scale transformation, graying and background estimation; then, a deep learning network is built, and an image obtained through image preprocessing operation is input into the deep learning network to obtain a binarized image; and finally, constructing an electronic paper display system, and carrying out mapping coding on the binarized image obtained in the deep learning step. According to the invention, by designing the degraded document image mapping method for electronic paper display, the image mapping performance is improved, and the display effect and the readability of the electronic paper on the degraded document image are improved.
Description
Technical Field
The invention relates to the technical field of display, in particular to a degraded document image mapping method for electronic paper display.
Background
With the rapid development of information technology, electronic reading has become a great trend. Electronic paper is used as a novel display device, adopts black-and-white binary images to display text images, and is widely applied to electronic book reading. The electronic paper has the characteristic of bistable state, can keep stable state for a long time without consuming electricity after being powered off, has the display effect similar to that of common newspapers, and therefore, the electronic paper has the possibility of replacing the traditional paper newspapers and becomes a new generation of mass reading media. Currently, in electronic paper display applications, the processing of degraded document images is a challenging problem. The degraded document image refers to a document image in which the image quality is degraded due to aging, offset, damage, and the like of paper. The images are characterized by more noise and distortion, less color information, and different degrees of spatial and temporal correlation. Such images are directly displayed on the electronic paper, which results in poor display effect and influences the reading experience of the user.
Disclosure of Invention
In view of the above, the present invention aims to provide a degraded document image mapping method for electronic paper display, which solves the problem of poor display effect of degraded document images on electronic paper by constructing an electronic paper display system and applying a binary algorithm based on deep learning to electronic paper display.
In order to achieve the above purpose, the invention adopts the following technical scheme: a degraded document image mapping method for electronic paper display comprises the following steps:
step S1, an electronic paper display hardware platform is built, wherein the electronic paper display hardware platform comprises a main control chip, an electronic paper driving board and an electronic paper display screen;
s2, constructing an image processing system, and performing image preprocessing operation on the degraded document image, wherein the image preprocessing operation comprises scale transformation, graying and background estimation;
s3, setting up a deep learning network, inputting the image obtained by the image preprocessing operation in the step S2 into the deep learning network to obtain a binary image, wherein the deep learning network comprises a U-Net++ network and an improved double-attention mechanism CBMA module;
and S4, building an electronic paper display system, and carrying out mapping coding on the binary image obtained in the step S3.
In a preferred embodiment: the step S1 is specifically implemented as follows:
step S11, the electronic paper can normally operate by the connecting device;
step S12, communicating with the electronic paper driving board through an SPI interface, and sending the image code to be displayed to the electronic paper driving board for processing;
and step S13, calling the driving waveform by the driving chip to display the image data processed in the step S12 on the electronic paper display screen.
In a preferred embodiment: the step S2 is specifically implemented as follows:
step S21, uniformly scaling the size of the input picture into the display resolution of the electronic paper by adopting an equal-ratio scaling method, namely performing scale conversion;
step S22, carrying out graying on the image subjected to the scale conversion in the step S21 by adopting a weighted average method;
and S23, performing stroke width conversion on the stroke width of the characters in the scale conversion grayscale image obtained in the step S22 to obtain the stroke width, performing background estimation through morphological closing operation to obtain a background image, and finally performing difference inversion on the scale conversion grayscale image obtained in the step S22 and the background image to obtain a foreground image.
In a preferred embodiment: the step S3 is specifically implemented as follows:
step S31, replacing an original channel attention mechanism in a dual attention mechanism CBMA module with a selective kernel network SKNet model to form an improved dual attention mechanism CBMA module;
step S32, the module obtained in the step S31 is applied to the end of the U-Net++ network to form an overall deep learning network architecture;
step S33, converting the gray level image output by the network architecture of step S32 into a binary image.
In a preferred embodiment: the step S4 is specifically implemented as follows:
s41, mapping and encoding the result obtained in the step S3 according to the mapping relation from the binary image to the display and encoding;
step S42, initializing a general input/output port of the main control chip, setting an enabling power supply of the driving chip according to the use condition of an internal/external power supply, configuring an internal register of the driving chip to enable driving to be electrified, driving the electronic paper to complete the initialization of erasing to full white and driving display, and storing display codes into the main control chip;
and S43, the driving chip calls the display code in the main control chip to drive the electronic paper screen to display the binarized image of the degraded document image.
In a preferred embodiment: the main control chip judges whether to switch the image when the electronic paper is in a dormant state, if the image is switched, the electronic paper is initialized, the driving chip recalls the display code in the main control chip and drives the electronic paper screen to display the binary image of the degraded document image again; and if the image is not switched, keeping the electronic paper in a dormant state.
In a preferred embodiment: the deep learning network comprises the following steps:
(1) Performing convolution and downsampling operation on an input image to form an encoder;
(2) An identity mapping crossing some layers of the encoder, adding the result of the identity mapping and the output of the layers of the encoder, and completing jump connection;
(3) Convoluting and downsampling the jump connection result to form a decoder;
(4) The output of the decoder is subjected to a dual attention mechanism and a convolution operation to obtain the final output image.
Compared with the prior art, the invention has the following beneficial effects: according to the invention, by designing the degraded document image mapping method for electronic paper display, the image mapping performance is improved, and the display effect and the readability of the electronic paper on the degraded document image are improved.
Drawings
FIG. 1 is an overall flowchart of a degraded document image mapping method for electronic paper display according to a preferred embodiment of the present invention;
FIG. 2 is a deep learning network model framework constructed in accordance with a preferred embodiment of the present invention.
Detailed Description
The invention will be further described with reference to the accompanying drawings and examples.
It should be noted that the following detailed description is illustrative and is intended to provide further explanation of the present application. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.
It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of example embodiments in accordance with the present application; as used herein, the singular is also intended to include the plural unless the context clearly indicates otherwise, and furthermore, it is to be understood that the terms "comprises" and/or "comprising" when used in this specification are taken to specify the presence of stated features, steps, operations, devices, components, and/or combinations thereof.
A degraded document image mapping method for electronic paper display is disclosed, and reference is made to FIGS. 1-2. The method comprises the following steps:
step S1, an electronic paper display hardware platform is built, wherein the electronic paper display hardware platform comprises a main control chip, an electronic paper driving board and an electronic paper display screen;
s2, constructing an image processing system, and performing image preprocessing operation including scale transformation and background estimation;
s3, setting up a deep learning network, and adding a CBMA improved dual attention module on the basis of U-Net++ to perform binarization processing;
and S4, building an electronic paper display system, and carrying out mapping coding on the image after document binarization.
In this embodiment, step S1 is to build an electronic paper display hardware platform, that is, connect all devices to enable the electronic paper to operate normally, where the electronic paper display hardware platform includes a main control chip, an electronic paper driving board and an electronic paper display screen, and specifically includes the following steps:
step S11, connecting a main control chip, an electronic paper driving board and an electronic paper display screen to enable the electronic paper to normally operate;
step S12, communicating with the electronic paper driving board through an SPI interface, and sending the image code to be displayed to the electronic paper driving board for processing;
and step S13, calling the driving waveform by the driving chip to display the image data processed in the step S12 on the electronic paper display screen.
Next, step S2 is to build an image processing system, and perform an image preprocessing operation on the degraded document image, where the image preprocessing operation includes scale transformation, graying and background estimation, and specifically includes the following steps:
and S21, uniformly scaling the size of the input picture into the display resolution of the electronic paper by adopting an equal-ratio scaling method, namely performing scale conversion. The resolution of the electronic paper used is 800×480, and the specific operation is as follows: if the aspect ratio of the image is greater than or equal to 0.6, limiting the height, scaling the image to an image with the height of 480 in an equal ratio, otherwise limiting the width, scaling the image to an image with the width of 800 in an equal ratio, adding blanks on the upper side and the lower side, supplementing the whole resolution of the image to 800 multiplied by 480, and enabling the image to be positioned in the center of the electronic paper display screen.
Step S22, carrying out graying on the image subjected to the scale conversion in the step S21 by adopting a weighted average method;
step S23, performing stroke width conversion on the stroke width of the characters in the scale conversion gray-scale image obtained in step S22 to obtain the stroke width, and communicatingAnd (3) performing background estimation through morphological closing operation to obtain a background image, and finally performing difference inversion on the scale transformation gray-scale image obtained in the step (S22) and the background image to obtain a foreground image. The original size of the selected structure is (2S W +1)×(2S W +1), wherein S W Representing the width of a text stroke.
Further, the step S3 is to build a deep learning network, the image obtained by the image preprocessing operation in the step S2 is input into the deep learning network to obtain a binary image, and the deep learning network comprises a U-Net++ network and an improved double-attention mechanism CBMA module, and specifically comprises the following steps:
step S31, replacing an original channel attention mechanism in a dual attention mechanism CBMA module with a selective kernel network SKNet model to form an improved dual attention module;
step S32, the module obtained in the step S31 is applied to the end of the U-Net++ network to form an overall deep learning network architecture;
step S33, converting the gray level image output by the network architecture of step S32 into a binary image.
Further, step S4, building an electronic paper display system, and mapping and encoding the binarized image obtained in step S3, specifically including the following steps:
step S41, according to the mapping relation from the binary image to the display code, mapping and coding the result obtained in the step S3, wherein the specific algorithm is as follows: each pixel point of the image is traversed. If the pixel value is 255, the point position is 0, if the pixel point is 0, the point position is 1, then 8 pixel points are a group, and are regarded as 8-bit binary numbers, weighted addition is carried out to obtain a 48000 decimal one-dimensional array, and then the decimal is converted into 16 decimal to obtain a final code;
step S42, initializing a general input/output port of the main control chip, setting an enabling power supply of the driving chip according to the use condition of an internal/external power supply, configuring an internal register of the driving chip to enable driving to be electrified, driving the electronic paper to complete the initialization of erasing to full white and driving display, and storing display codes into the main control chip;
and S43, the driving chip calls the display code in the main control chip to drive the electronic paper screen to display the binarized image of the degraded document image.
The main control chip judges whether to switch the image when the electronic paper is in a dormant state, if the image is switched, the electronic paper is initialized, the driving chip recalls the display code in the main control chip and drives the electronic paper screen to display the binary image of the degraded document image again; and if the image is not switched, keeping the electronic paper in a dormant state.
The deep learning network comprises the following steps:
(1) Performing convolution and downsampling operation on an input image to form an encoder;
(2) An identity mapping crossing some layers of the encoder, adding the result of the identity mapping and the output of the layers of the encoder, and completing jump connection;
(3) Convoluting and downsampling the jump connection result to form a decoder;
(4) The output of the decoder is subjected to a dual attention mechanism and a convolution operation to obtain the final output image.
The above is a preferred embodiment of the present invention, and all changes made according to the technical solution of the present invention belong to the protection scope of the present invention when the generated functional effects do not exceed the scope of the technical solution of the present invention.
Claims (7)
1. The degraded document image mapping method for electronic paper display is characterized by comprising the following steps of:
step S1, an electronic paper display hardware platform is built, wherein the electronic paper display hardware platform comprises a main control chip, an electronic paper driving board and an electronic paper display screen;
s2, constructing an image processing system, and performing image preprocessing operation on the degraded document image, wherein the image preprocessing operation comprises scale transformation, graying and background estimation;
s3, setting up a deep learning network, inputting the image obtained by the image preprocessing operation in the step S2 into the deep learning network to obtain a binary image, wherein the deep learning network comprises a U-Net++ network and an improved double-attention mechanism CBMA module;
and S4, building an electronic paper display system, and carrying out mapping coding on the binary image obtained in the step S3.
2. The method for mapping a degraded document image displayed on electronic paper according to claim 1, wherein the method comprises the steps of: the step S1 is specifically implemented as follows:
step S11, the electronic paper can normally operate by the connecting device;
step S12, communicating with the electronic paper driving board through an SPI interface, and sending the image code to be displayed to the electronic paper driving board for processing;
and step S13, calling the driving waveform by the driving chip to display the image data processed in the step S12 on the electronic paper display screen.
3. The method for mapping a degraded document image displayed on electronic paper according to claim 1, wherein the method comprises the steps of: the step S2 is specifically implemented as follows:
step S21, uniformly scaling the size of the input picture into the display resolution of the electronic paper by adopting an equal-ratio scaling method, namely performing scale conversion;
step S22, carrying out graying on the image subjected to the scale conversion in the step S21 by adopting a weighted average method;
and S23, performing stroke width conversion on the stroke width of the characters in the scale conversion grayscale image obtained in the step S22 to obtain the stroke width, performing background estimation through morphological closing operation to obtain a background image, and finally performing difference inversion on the scale conversion grayscale image obtained in the step S22 and the background image to obtain a foreground image.
4. The method for mapping a degraded document image displayed on electronic paper according to claim 1, wherein the method comprises the steps of: the step S3 is specifically implemented as follows:
step S31, replacing an original channel attention mechanism in a dual attention mechanism CBMA module with a selective kernel network SKNet model to form an improved dual attention mechanism CBMA module;
step S32, the module obtained in the step S31 is applied to the end of the U-Net++ network to form an overall deep learning network architecture;
step S33, converting the gray level image output by the network architecture of step S32 into a binary image.
5. The method for mapping a degraded document image displayed on electronic paper according to claim 1, wherein the method comprises the steps of: the step S4 is specifically implemented as follows:
s41, mapping and encoding the result obtained in the step S3 according to the mapping relation from the binary image to the display and encoding;
step S42, initializing a general input/output port of the main control chip, setting an enabling power supply of the driving chip according to the use condition of an internal/external power supply, configuring an internal register of the driving chip to enable driving to be electrified, driving the electronic paper to complete the initialization of erasing to full white and driving display, and storing display codes into the main control chip;
and S43, the driving chip calls the display code in the main control chip to drive the electronic paper screen to display the binarized image of the degraded document image.
6. The method for mapping a degraded document image displayed on electronic paper according to claim 1, wherein the method comprises the steps of: the main control chip judges whether to switch the image when the electronic paper is in a dormant state, if the image is switched, the electronic paper is initialized, the driving chip recalls the display code in the main control chip and drives the electronic paper screen to display the binary image of the degraded document image again; and if the image is not switched, keeping the electronic paper in a dormant state.
7. The method for mapping a degraded document image displayed on electronic paper according to claim 1, wherein the method comprises the steps of: the deep learning network comprises the following steps:
(1) Performing convolution and downsampling operation on an input image to form an encoder;
(2) An identity mapping crossing some layers of the encoder, adding the result of the identity mapping and the output of the layers of the encoder, and completing jump connection;
(3) Convoluting and downsampling the jump connection result to form a decoder;
(4) The output of the decoder is subjected to a dual attention mechanism and a convolution operation to obtain the final output image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311599318.1A CN117632072A (en) | 2023-11-28 | 2023-11-28 | Electronic paper display-oriented degraded document image mapping method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311599318.1A CN117632072A (en) | 2023-11-28 | 2023-11-28 | Electronic paper display-oriented degraded document image mapping method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117632072A true CN117632072A (en) | 2024-03-01 |
Family
ID=90035077
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311599318.1A Pending CN117632072A (en) | 2023-11-28 | 2023-11-28 | Electronic paper display-oriented degraded document image mapping method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117632072A (en) |
-
2023
- 2023-11-28 CN CN202311599318.1A patent/CN117632072A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109472270B (en) | Image style conversion method, device and equipment | |
| US6781600B2 (en) | Shape processor | |
| Douak et al. | Color image compression algorithm based on the DCT transform combined to an adaptive block scanning | |
| CN104838653A (en) | Lossless image compression using differential transfer | |
| US11887280B2 (en) | Method, system, and computer-readable medium for improving quality of low-light images | |
| CN112929624B (en) | Encoding method, encoding device, electronic equipment and computer readable storage medium | |
| CN112036137A (en) | Deep learning-based multi-style calligraphy digital ink simulation method and system | |
| Han et al. | Vector quantizer based block truncation coding for color image compression in LCD overdrive | |
| US9781419B2 (en) | Compressing image data | |
| CN117632072A (en) | Electronic paper display-oriented degraded document image mapping method | |
| WO2008008131A1 (en) | Providing multiple and native representations of an image | |
| Cai et al. | Beautified QR code with high storage capacity using sequential module modulation | |
| CN112887722A (en) | Lossless image compression method | |
| Zheng et al. | A novel gray image representation using overlapping rectangular NAM and extended shading approach | |
| Said | Compression of compound images and video for enabling rich media in embedded systems | |
| JP2005055825A (en) | Image display device, image display method and image display program | |
| CN115909378A (en) | Document text detection model training method and document text detection method | |
| US20080226164A1 (en) | Image data decoding device, image data encoding device, image processing device, and electronic instrument | |
| US11412260B2 (en) | Geometric transforms for image compression | |
| CN113038134A (en) | Picture processing method, intelligent terminal and storage medium | |
| CN113034625B (en) | Lossless compression method based on picture, intelligent terminal and storage medium | |
| CN116996695B (en) | Panoramic image compression method, device, equipment and medium | |
| CN115631115B (en) | Dynamic image restoration method based on recursion transform | |
| Abebe et al. | Deep learning approaches for whiteboard image quality enhancement | |
| JP2013186826A (en) | Information code creation device, information code reading device and information code |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |