CN112435186A - Fingerprint image enhancement method based on double-rule matrix direction field - Google Patents
Fingerprint image enhancement method based on double-rule matrix direction field Download PDFInfo
- Publication number
- CN112435186A CN112435186A CN202011315342.4A CN202011315342A CN112435186A CN 112435186 A CN112435186 A CN 112435186A CN 202011315342 A CN202011315342 A CN 202011315342A CN 112435186 A CN112435186 A CN 112435186A
- Authority
- CN
- China
- Prior art keywords
- matrix
- fingerprint
- direction field
- fingerprint image
- double
- 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.)
- Granted
Links
- 239000011159 matrix material Substances 0.000 title claims abstract description 71
- 238000000034 method Methods 0.000 title claims abstract description 24
- 230000000877 morphologic effect Effects 0.000 claims abstract description 8
- 230000009977 dual effect Effects 0.000 claims description 8
- 238000001914 filtration Methods 0.000 claims description 8
- 238000005516 engineering process Methods 0.000 description 4
- 230000011218 segmentation Effects 0.000 description 3
- 230000010339 dilation Effects 0.000 description 2
- 230000002708 enhancing effect Effects 0.000 description 2
- 230000003628 erosive effect Effects 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 230000002093 peripheral effect Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 238000011840 criminal investigation Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000000513 principal component analysis Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Images
Classifications
-
- G06T5/77—
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/20—Image enhancement or restoration by the use of local operators
- G06T5/30—Erosion or dilatation, e.g. thinning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/194—Segmentation; Edge detection involving foreground-background segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/12—Fingerprints or palmprints
- G06V40/1347—Preprocessing; Feature extraction
- G06V40/1359—Extracting features related to ridge properties; Determining the fingerprint type, e.g. whorl or loop
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20021—Dividing image into blocks, subimages or windows
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20172—Image enhancement details
Abstract
The invention discloses a fingerprint image enhancement method based on a double-rule matrix direction field, which comprises the following steps: firstly, judging whether a point is on a fingerprint by calculating the mean value and the variance of an image to obtain a foreground and a background; designing a double-rule matrix, acquiring fingerprint direction field characteristics through the double-rule matrix, calculating a direction characteristic formula to judge whether the foreground area pixel points are positioned on the ridge line, if so, setting the foreground area pixel points to be black, otherwise, setting the foreground area pixel points to be white; removing burrs and cavities in the fingerprint graph according to the neighborhood of each pixel point; and finally, performing morphological operation to obtain a final fingerprint image, thereby realizing the image enhancement of the fingerprint. The invention realizes the fingerprint image enhancement method based on the direction field, and realizes the fingerprint image enhancement by providing a double square method to determine the direction field characteristic area and the direction characteristic formula.
Description
Technical Field
The invention belongs to the field of image processing, and particularly relates to a fingerprint image enhancement method based on a double-rule matrix direction field in the field of images.
Background
The fingerprint identification technology is one of the highest occupation rate technologies in the biological identification technology, and is widely applied to the fields of access control technology, criminal investigation, fingerprint payment and the like. In a criminal incident, a criminal suspect often leaves a fingerprint at a criminal scene, and the suspect can be determined by comparing the fingerprint at the scene with the fingerprint in a database, however, the fingerprint at the scene is often blurred, so that difficulty is increased for subsequent fingerprint comparison, and the fingerprint needs to be enhanced.
The current fingerprint image enhancement algorithm mainly comprises histogram equalization, principal component analysis, Gabor filtering, median filtering, Sobel filtering and the like, however, the method can also cause noise enhancement while enhancing, and causes the fingerprint ridge line profile to be fuzzy; the deep learning algorithm is used for enhancing the fingerprints, such as a convolutional neural network and a confrontation generation network, so that the time complexity of the fingerprint processing algorithm is high, the requirement on hardware is high, and the fingerprints are difficult to process on a common computer; in addition, if water, oil, or the like is present on the finger, the fingerprint image is blurred or broken, and is distinguished from the actual fingerprint texture.
Disclosure of Invention
The invention aims to provide a fingerprint image enhancement method based on a double-rule matrix direction field, which improves the enhanced fingerprint feature details and removes redundant noise, has low time complexity and meets the application requirement of a basic computer.
In order to achieve the purpose of the invention, the technical scheme adopted by the invention is that the fingerprint image enhancement method based on the direction field of the dual-rule matrix comprises the following steps:
step 1: acquiring a fingerprint image, and determining an area where the fingerprint is located as a foreground area;
step 2: designing a double-rule matrix, and acquiring fingerprint direction field characteristics through the double-rule matrix;
and step 3: judging whether foreground pixel points are located on ridge lines or not through a direction field characteristic formula;
and 4, step 4: deleting flaws in the image output in the step 3;
and 5: and (4) performing morphological filtering on the image processed in the step (4) and outputting an enhanced fingerprint image.
Preferably, the step 1 comprises the steps of:
step 1.1: acquiring a fingerprint image, converting the fingerprint image into a gray image, establishing a coordinate system for the gray image, and calculating a pixel MEAN and a variance VAR of the gray image;
step 1.2: carrying out normalization processing on pixel points of coordinates (x, y) of the gray level image;
step 1.3: dividing the processed gray level image into preset image blocks, judging the pixel MEAN value of each image block, if the pixel MEAN value is smaller than MEAN, the current image block is a fingerprint area and is a foreground, otherwise, the current image block is a background;
step 1.4: and (4) repeating the step 1.3 until all the image blocks are traversed, taking all the foregrounds as the foreground areas of the fingerprint images, and taking the rest parts as the background areas.
Preferably, in step 1.1, the pixel mean valueVariance (variance)Wherein M and N are the number of pixels of the gray image in the length and width directions, respectively, and G (x, y) represents the pixel value of a pixel point of coordinates (x, y);
in said step 1.2, toNormalizing pixel points of coordinates (x, y) of the gray image, wherein M0And VAR0Respectively, the desired mean and the desired variance.
Preferably, the dual rule matrix in step 2 includes an outer matrix and an inner matrix.
Preferably, the external matrix is a sparse interval matrix a x a, where a is the number of pixels on the edge of the external matrix; the internal matrix is a continuous and continuous non-interval matrix b, and b is the number of pixel points on the edge of the internal matrix.
Preferably, b is an odd number, and a-2 × b-1.
Specifically, the step 2 includes the steps of:
step 2.1: determining double-rule matrixes a and b, wherein the periphery of the double-rule matrixes is a sparse interval matrix a, and a is the number of pixel points on the edge of the peripheral matrix; the internal matrix is a continuous and continuous non-interval matrix b, wherein b is the number of pixel points on the edge of the internal matrix; with the proviso that b is odd and a-2 x b-1; in the specification, b is 7, and a is 13;
step 2.2: the process of confirming the direction field area by the dual-rule matrix is as follows, each pixel point C of the foreground area is made into a 13-by-13 square matrix by taking C as the center; calibrating 1A to 12A, 1B to 12B, 1A 'to 12A' and 1B 'to 12B' in a matrix according to a preset rule;
preferably, in step 2.2, 1A to 12A and 1A 'to 12A' are sequentially marked on the side of the rectangle with C as the center point, the side length of the internal rectangle is 7 pixel points, and 1A 'to 12A' and 1A to 12A are symmetrical about the center of C.
Preferably, in step 2.2, 1B to 12B, 1B 'to 12B' are sequentially marked on the edge of the rectangle with C as the center point at intervals of 1 pixel point, the length of the outer rectangle is 13 pixel points, and 1B 'to 12B' and 1B to 12B are symmetrical about C center.
Preferably, in step 3, the central point C of the dual-rule matrix is respectively connected to the pixels of the dual-rule matrix, and the sum of the pixel values of all the lines in the direction n is respectively calculated, where the maximum value is assigned as SmaxMinimum value of SminThe average value in all directions is Save(ii) a If the formula S of the direction field characteristic is satisfiedmax+Smin+4*c>3*SaveThen point C lies on the valley line and is assigned 255, i.e. white, otherwise point C lies on the ridge line.
Specifically, the step 3 includes the steps of:
step 3.1: taking the lines passing through nA, nB, nA 'and nB' as the direction n, respectively calculating the pixel value sum of all the lines in the direction n, and assigning S to the maximum valuemaxMinimum value assignment SminMean value assignment S in all directionsave;
Step 3.2: judging the position of the pixel point by a direction field characteristic formula, namely if S is satisfiedmax+Smin+4*c>3*SaveThen point C lies on the valley line and is assigned 255, i.e. white, otherwise point C lies on the ridge line.
Preferably, in step 4, the flaws include burrs and voids.
Preferably, if any pixel is black and there are three or more white pixels in the four-adjacent domain of the pixel, the pixel is an end point of a burr or a ridge, and the current pixel is assigned to be 255, that is, white.
Preferably, if any pixel is white and there are three or more black pixels in the four-adjacent domain of the pixel, the pixel is a hole, and the current pixel is assigned to be 0, that is, black.
Preferably, the step 5 comprises the steps of:
step 5.1: scanning each pixel in the image output in the step 3 by using a structural element of 3x3 by using a morphological erosion operation, and performing an AND operation on the structural element and the area of the image covered by the structural element, wherein if the result is 1, the pixel value is 1;
step 5.2: using the dilation operation in morphology, each pixel of the image processed in step 4.1 is scanned with a structuring element of 3 × 3, and the structuring element and the area of the image it covers are anded together, and if the result is 0, the pixel value is 0.
The invention provides an optimized fingerprint image enhancement method based on a double-rule matrix direction field, which has the beneficial effects that:
(1) whether the point is positioned on a ridge line is judged according to the average gray value of each pixel point in 12 directions to enhance the fingerprint image, and meanwhile, the influence of burrs and holes on the fingerprint image is removed;
(2) for sparse fingerprint images, the ridge line distance is larger due to factors such as the degree of pressing, and the like, whether pixel points are located on ridge lines or not is predicted through 12 directions of the sparse fingerprint images, so that partial sparse fingerprints are connected;
(3) for a compact fingerprint image, other algorithms do not reserve the segmentation details of the fingerprint detail part, so that ridges present aggregative property and fuzziness, and the method can well extract fingerprint characteristics.
Drawings
FIG. 1 is a flow chart of the present invention;
fig. 2 is a schematic diagram of a dual rule matrix of the present invention, wherein, taking B as 7 and a as 13 as an example, solid dots are used to identify 1A to 12A, 1B to 12B, 1A 'to 12A', 1B 'to 12B', and arrows indicate a direction 1 to a direction 12;
fig. 3 shows the effect of the present invention, which is 3 groups in total, wherein in any one group, (a) is the original image, (b) is the normalized image, (c) is the fingerprint area image, (d) is the direction field ridge segmentation image, (e) is the fingerprint image after removing the holes and burrs, and (f) is the result image after morphological filtering.
Detailed Description
The present invention is described in further detail with reference to the following examples, but the scope of the present invention is not limited thereto.
The invention provides a fingerprint image enhancement method based on a double rule matrix direction field, which judges whether a point is positioned on a ridge line or not by estimating 12 direction values of a fingerprint so as to realize the segmentation of the fingerprint.
To achieve the above object, the method of the present invention comprises the following steps.
Step 1: and acquiring a fingerprint image, and determining the area where the fingerprint is located as a foreground area.
The step 1 comprises the following steps:
step 1.1: acquiring a fingerprint image, converting the fingerprint image into a gray image, establishing a coordinate system for the gray image, and calculating a pixel MEAN and a variance VAR of the gray image;
in said step 1.1, the pixelMean valueVariance (variance)Wherein M and N are the number of pixels of the gray image in the length and width directions, respectively, and G (x, y) represents the pixel value of a pixel point of coordinates (x, y);
step 1.2: carrying out normalization processing on pixel points of coordinates (x, y) of the gray level image;
in said step 1.2, toNormalizing pixel points of coordinates (x, y) of the gray image, wherein M0And VAR0The expected mean and the expected variance are assigned values of 150 and 120, respectively;
step 1.3: dividing the processed gray-scale image into preset image blocks, generally 5 × 5 image blocks, judging the pixel MEAN value of each image block, if the pixel MEAN value is less than MEAN, the current image block is a fingerprint area and is a foreground, otherwise, the current image block is a background;
step 1.4: and (4) repeating the step 1.3 until all the image blocks are traversed, taking all the foregrounds as the foreground areas of the fingerprint images, and taking the rest parts as the background areas.
Step 2: and designing a double-rule matrix, and acquiring the fingerprint direction field characteristics through the double-rule matrix.
The dual rule matrix in the step 2 comprises an outer matrix and an inner matrix.
The external matrix is a sparse interval matrix a, wherein a is the number of pixel points on the edge of the external matrix; the internal matrix is a continuous and continuous non-interval matrix b, and b is the number of pixel points on the edge of the internal matrix.
b is an odd number, and a-2 b-1.
The step 2 comprises the following steps:
step 2.1: determining double-rule matrixes a and b, wherein the periphery of the double-rule matrixes is a sparse interval matrix a, and a is the number of pixel points on the edge of the peripheral matrix; the internal matrix is a continuous and continuous non-interval matrix b, wherein b is the number of pixel points on the edge of the internal matrix; with the proviso that b is odd and a-2 x b-1; in the specification, b is 7, and a is 13;
step 2.2: the process of confirming the direction field area by the dual-rule matrix is as follows, each pixel point C of the foreground area is made into a 13-by-13 square matrix by taking C as the center; calibrating 1A to 12A, 1B to 12B, 1A 'to 12A' and 1B 'to 12B' in a matrix according to a preset rule;
in step 2.2, 1A to 12A and 1A 'to 12A' are sequentially calibrated on the side of the rectangle with C as the center point, the side length of the internal rectangle is 7 pixel points, and 1A 'to 12A' and 1A to 12A are symmetrical with C center.
In step 2.2, 1B to 12B, 1B 'to 12B' are sequentially marked on the edge of the rectangle with C as the center point by 1 pixel point at intervals, the side length of the external rectangle is 13 pixel points, and 1B 'to 12B' and 1B to 12B are symmetrical about the center of C.
In the present invention, 1A to 12A, 1B to 12B, 1A 'to 12A' and 1B 'to 12B' are all calibrated clockwise or counterclockwise at the same time.
In the invention, taking C as the center and making 13 × 13 square matrixes as an optimal scheme of comprehensive operation amount and precision, a person skilled in the art can set the size, the number of directions and corresponding calibration rules of the square matrixes according to requirements.
And step 3: and judging whether the foreground pixel points are positioned on the ridge line or not through a direction field characteristic formula.
In the step 3, the central point C of the dual-rule matrix is respectively connected with the pixel points of the dual-rule matrix, and the sum of the pixel values of all the lines in the direction n is respectively calculated, wherein the maximum value is assigned as SmaxMinimum value of SminThe average value in all directions is Save(ii) a If the formula of the direction field characteristic is satisfied
Smax+Smin+4*c>3*SaveThen point C lies on the valley line and is assigned 255, i.e. white, otherwise point C lies on the ridge line.
The step 3 comprises the following steps:
step 3.1: to pass throughThe lines of nA, nB, nA 'and nB' are taken as the direction n, the pixel value sum of all the lines in the direction n is respectively calculated, and the maximum value is assigned to SmaxMinimum value assignment SminMean value assignment S in all directionsave;
Step 3.2: judging the position of the pixel point by a direction field characteristic formula, namely if S is satisfiedmax+Smin+4*c>3*SaveThen point C lies on the valley line and is assigned 255, i.e. white, otherwise point C lies on the ridge line.
In the present invention, it is obvious that C in the directional field characteristic formula is the pixel value of the center point C.
And 4, step 4: and (4) deleting the flaws in the image output in the step (3).
In step 4, the defects include burrs and voids.
If any pixel is black and the four adjacent domains of the pixel have three or more white pixels, the pixel is the end point of the burr or the ridge line, and the current pixel is assigned to be 255, namely white.
If any pixel point is white and the four adjacent domains of the pixel point have three or more than three black pixel points, the pixel point is a hole, and the current pixel point is assigned to be 0, namely black.
In the invention, when processing the burr, a condition exists, namely, besides the burr, three white pixel points also exist in four adjacent domains of the end point part of the ridge line of the fingerprint, but generally, setting a black pixel point of the end point of the ridge line of the fingerprint to be white does not influence the overall presentation of the fingerprint.
And 5: and (4) performing morphological filtering on the image processed in the step (4) and outputting an enhanced fingerprint image.
The step 5 comprises the following steps:
step 5.1: scanning each pixel in the image output in the step 3 by using a structural element of 3x3 by using a morphological erosion operation, and performing an AND operation on the structural element and the area of the image covered by the structural element, wherein if the result is 1, the pixel value is 1;
step 5.2: using the dilation operation in morphology, each pixel of the image processed in step 4.1 is scanned with a structuring element of 3 × 3, and the structuring element and the area of the image it covers are anded together, and if the result is 0, the pixel value is 0.
In the invention, in the morphological filtering process, the structural elements of 3 × 3 are not unique, and those skilled in the art can set different structural elements to process the output image according to the requirements.
Claims (5)
1. A fingerprint image enhancement method based on a double rule matrix direction field is characterized in that: the method comprises the following steps:
step 1: acquiring a fingerprint image, and determining an area where the fingerprint is located as a foreground area;
step 2: designing a double-rule matrix, and acquiring fingerprint direction field characteristics through the double-rule matrix;
and step 3: judging whether foreground pixel points are located on ridge lines or not through a direction field characteristic formula;
and 4, step 4: deleting flaws in the image output in the step 3;
and 5: and (4) performing morphological filtering on the image processed in the step (4) and outputting an enhanced fingerprint image.
2. The fingerprint image enhancement method based on the direction field of the dual rule matrix as claimed in claim 1, wherein: the dual rule matrix in the step 2 comprises an outer matrix and an inner matrix.
3. The fingerprint image enhancement method based on the direction field of the dual rule matrix as claimed in claim 2, wherein: the external matrix is a sparse interval matrix a, wherein a is the number of pixel points on the edge of the external matrix; the internal matrix is a continuous and continuous non-interval matrix b, and b is the number of pixel points on the edge of the internal matrix.
4. The fingerprint image enhancement method based on the direction field of the dual rule matrix as claimed in claim 3, wherein: b is an odd number, and a-2 b-1.
5. The fingerprint image enhancement method based on the direction field of the dual rule matrix as claimed in claim 1, wherein: in the step 3, the central point C of the dual-rule matrix is respectively connected with the pixel points of the dual-rule matrix, and the sum of the pixel values of all the lines in the direction n is respectively calculated, wherein the maximum value is assigned as SmaxMinimum value of SminThe average value in all directions is Save(ii) a If the formula S of the direction field characteristic is satisfiedmax+Smin+4*c>3*SaveThen point C lies on the valley line and is assigned 255, i.e. white, otherwise point C lies on the ridge line.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011315342.4A CN112435186B (en) | 2020-11-20 | Fingerprint image enhancement method based on double-rule matrix direction field |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011315342.4A CN112435186B (en) | 2020-11-20 | Fingerprint image enhancement method based on double-rule matrix direction field |
Publications (2)
Publication Number | Publication Date |
---|---|
CN112435186A true CN112435186A (en) | 2021-03-02 |
CN112435186B CN112435186B (en) | 2024-04-26 |
Family
ID=
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116823679A (en) * | 2023-08-30 | 2023-09-29 | 山东龙腾控股有限公司 | Full-automatic fingerprint lock fingerprint image enhancement method based on artificial intelligence |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1421815A (en) * | 2001-11-29 | 2003-06-04 | 田捷 | Fingerprint image enhancement method based on knowledge |
CN101079102A (en) * | 2007-06-28 | 2007-11-28 | 中南大学 | Fingerprint identification method based on statistic method |
CN102103692A (en) * | 2011-03-17 | 2011-06-22 | 电子科技大学 | Fingerprint image enhancing method |
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1421815A (en) * | 2001-11-29 | 2003-06-04 | 田捷 | Fingerprint image enhancement method based on knowledge |
CN101079102A (en) * | 2007-06-28 | 2007-11-28 | 中南大学 | Fingerprint identification method based on statistic method |
CN102103692A (en) * | 2011-03-17 | 2011-06-22 | 电子科技大学 | Fingerprint image enhancing method |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116823679A (en) * | 2023-08-30 | 2023-09-29 | 山东龙腾控股有限公司 | Full-automatic fingerprint lock fingerprint image enhancement method based on artificial intelligence |
CN116823679B (en) * | 2023-08-30 | 2023-12-05 | 山东龙腾控股有限公司 | Full-automatic fingerprint lock fingerprint image enhancement method based on artificial intelligence |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US7072523B2 (en) | System and method for fingerprint image enhancement using partitioned least-squared filters | |
US8712114B2 (en) | Elegant solutions for fingerprint image enhancement | |
CN112001302B (en) | Face recognition method based on face interesting region segmentation | |
US20110262013A1 (en) | Fingerprint matcher using iterative process and related methods | |
WO2017161636A1 (en) | Fingerprint-based terminal payment method and device | |
JP4901229B2 (en) | Red-eye detection method, apparatus, and program | |
CN114820625A (en) | Automobile top block defect detection method | |
Subasic et al. | Face image validation system | |
CN115272647A (en) | Lung image recognition processing method and system | |
CN112883941A (en) | Facial expression recognition method based on parallel neural network | |
JP4285640B2 (en) | Object identification method, apparatus and program | |
JP3020973B2 (en) | Image processing method | |
CN110349119B (en) | Pavement disease detection method and device based on edge detection neural network | |
CN112435186B (en) | Fingerprint image enhancement method based on double-rule matrix direction field | |
CN112435186A (en) | Fingerprint image enhancement method based on double-rule matrix direction field | |
CN116503871A (en) | Character segmentation preprocessing method, terminal device and computer readable storage medium | |
CN111986176B (en) | Crack image identification method, system, terminal and readable storage medium | |
CN113139577A (en) | Deep learning image classification method and system based on deformable convolution network | |
Prakram et al. | EVALUATION OF IMPROVED FUZZY INFERENCE SYSTEM TO PRESERVE IMAGE EDGE FOR IMAGE ANALYSIS. | |
JP2007524143A (en) | Palmprint authentication method | |
Okarma et al. | A Hybrid Method for Objective Quality Assessment of Binary Images | |
CN116245770B (en) | Endoscope image edge sharpness enhancement method and device, electronic equipment and storage medium | |
CN117557785B (en) | Image processing-based long-distance fishing boat plate recognition method | |
CN117423113B (en) | Adaptive denoising method for archive OCR (optical character recognition) image | |
Lin et al. | High-Accuracy Skew Estimation of Document Images. |
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 | ||
GR01 | Patent grant |