CN110415217A - Enhance preferred method based on subset guiding and the image set of the coefficient of variation - Google Patents
Enhance preferred method based on subset guiding and the image set of the coefficient of variation Download PDFInfo
- Publication number
- CN110415217A CN110415217A CN201910594023.2A CN201910594023A CN110415217A CN 110415217 A CN110415217 A CN 110415217A CN 201910594023 A CN201910594023 A CN 201910594023A CN 110415217 A CN110415217 A CN 110415217A
- Authority
- CN
- China
- Prior art keywords
- image set
- image
- algorithm
- subgraph
- variation
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- 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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30168—Image quality inspection
Landscapes
- Engineering & Computer Science (AREA)
- Quality & Reliability (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The invention proposes a kind of image sets based on subset guiding and the coefficient of variation to enhance preferred method, and mentioned method carries out grab sample to original image set first and obtains subgraph image set, represents original image set using subgraph image set.Then, sub- image set is enhanced using a variety of algorithm for image enhancement, obtain multiple enhanced subgraph image sets, a certain image quality evaluation criterion is reused to give a mark to multiple enhanced subgraph image sets one by one, according to marking as a result, calculating the average value and variance of each enhanced subgraph image set.Mentioned method calculates separately out the coefficient of variation of each enhanced subgraph image set according to average value and variance, and enhancing algorithm corresponding to the minimum coefficient of variation is the optimal algorithm of subgraph image set；According to the measurement system that subset guides, enhancing algorithm corresponding to the minimum coefficient of variation is also the optimal enhancing algorithm of original image set simultaneously.
Description
Technical field
The present invention provides a kind of preferred method based on subset guiding and the coefficient of variation for a variety of algorithm for image enhancement, belongs to
In image procossing and quality evaluation field.
Background technique
Image enhancement is using very extensive in people's daily life and production, it is to meet application-specific demand as mesh
, interested information in prominent image, usually as the preprocessing process of image analysis identification, the purpose is to adjust original image
Information so that it is more suitable for man-machine identification.To carry out the enhancing of robust to image, then need to find a kind of optimal image
Enhance algorithm.For single image, searching optimum image enhancing algorithm is relatively easy, i.e., using a variety of enhancing algorithms to single width figure
As being enhanced, several enhanced images are obtained, then enhanced image is beaten using image quality evaluation criterion
Point, enhancing algorithm is the optimal enhancing algorithm of the width image corresponding to the image of top score.
For the image set of great amount of images composition, it is in contrast relatively difficult to find optimal enhancing algorithm, a side
Face is that the quantity of enhancing algorithm is more, is on the other hand that image set is very big, operates very time-consuming.Existing image set enhancing is excellent
Choosing method is mainly averaging method, is enhanced first using a variety of enhancing algorithms image set, obtains multiple enhanced figures
Image set；Then every piece image in enhancing image set is given a mark and is averaged using image quality evaluation criterion, most
Enhancing algorithm corresponding to high average value is the optimal enhancing algorithm of the image set.However there is certain lack in averaging method
Fall into because averaging method only considers this index of average value, if in enhanced image set certain images mass fraction
Very big, the then easy fluctuation of average value is fluctuated, average value is difficult to represent the integral level of entire enhancing image set at this time.In addition, with
The arriving of big data era, image data is increasingly huge, and image set is usually very big, enhances whole image collection, then
It is given a mark to enhancing image set and to be averaged this process very time-consuming.
Summary of the invention
The object of the present invention is to provide a kind of reliabilities and the higher image set of efficiency to enhance preferred method, sieves for image set
Select a kind of algorithm for image enhancement of best performance.
In order to achieve the above object, the technical solution of the present invention is to provide a kind of based on subset guiding and the coefficient of variation
Image set enhances preferred method, which comprises the following steps:
Step 1: carrying out grab sample to original image set A and obtain subgraph image set B, and original image set A includes p width image,
Subgraph image set B includes n width image；
Step 2: a variety of algorithm for image enhancement (E are used1, E2..., Em) sub- image set B is enhanced, it obtains multiple
Enhanced subgraph image set (B1, B2..., Bm), m is the quantity of algorithm for image enhancement；
Step 3: using a certain kind image quality evaluation criterion Q successively to enhanced subgraph image set (B1, B2..., Bm)
It gives a mark, marking result is denoted as αij, wherein i be image label, i=1,2 ... n, j be various algorithm for image enhancement mark
Number, j=1,2 ... m；
Step 4: successively calculating enhances algorithm (E at each1, E2..., Em) under, αijAverage value UjAnd variance
Sj:
Step 5: it calculates separately out in each algorithm for image enhancement (E1, E2..., Em) under enhancer image set change
Different coefficient, the coefficient of variation of enhancer image set is defined as COV under jth kind algorithm for image enhancementj, then have:
Step 6: when obtaining each algorithm for image enhancement (E1, E2..., Em) under enhancer image set the coefficient of variation
Later, minimum coefficient of variation COV is picked outmin=min (COV1, COV2..., COVm), minimum coefficient of variation COVminIt is corresponding
Algorithm for image enhancement be the subgraph image set optimal enhancing algorithm；
Step 7: after the optimal enhancing algorithm for selecting subgraph image set B, according to the evaluation and test system that subset guides, subgraph image set B
Optimal enhancing algorithm be original image set A optimal enhancing algorithm.
In order to select the optimal algorithm of suitable original image set from a variety of algorithm for image enhancement, the invention proposes one kind
Enhance preferred method based on subset guiding and the image set of the coefficient of variation, by introducing the evaluation and test system of subset guiding, for super
Big original image set carries out grab sample to it and obtains subgraph image set, then carries out enhancing operation to sub- image set, imitates in this way
Rate greatly improves.Mentioned method not only allows for average value, it is also contemplated that variance has the characteristics that high reliability and efficient.
Detailed description of the invention
Fig. 1 is overall framework figure of the invention；
Fig. 2 is the specific flow chart for being the proposed method of the present invention.
Specific embodiment
With reference to the accompanying drawing, the present invention is further explained.It should be understood that these embodiments are merely to illustrate the present invention and do not have to
In limiting the scope of the invention.In addition, it should also be understood that, after reading the content taught by the present invention, those skilled in the art can be with
The present invention is made various changes or modifications, such equivalent forms equally fall within model defined by the application the appended claims
It encloses.
Fig. 1 is overall framework figure of the invention.The present embodiment illustrates proposed method by taking underwater picture collection as an example, selects
500 width underwater pictures as original image set A, choose histogram equalization, contrast-limited histogram equalization and
These three algorithm for image enhancement of the unsupervised color model of Iqbal et al. proposition are illustrated as example.Image quality evaluation
Criterion is UCIQE (Underwater Color Image Quality Evaluation) criterion, by Yang and Arcot et al.
It proposes, UCIQE is current most widely used underwater picture evaluating criterion of quality.Fig. 2 is be the proposed method of the present invention specific
Flow chart, specific implementation steps are as follows:
Step 1: 50 width images are randomly selected from 500 width original image set A, as subgraph image set B.
Step 2: histogram equalization (E is used1), contrast-limited histogram equalization (E2), unsupervised color model (E3)
Three kinds of algorithm for image enhancement enhance sub- image set B, obtain three enhanced subgraph image set (B1, B2, B3)。
Step 3: using UCIQE image quality evaluation criterion (Q) successively to enhanced subgraph image set (B1, B2, B3) into
Row marking, marking result are denoted as αij, wherein (i=1,2 ..., 50) be image label, scheme included in each subgraph collection i
As quantity is equal to 50；J (j=1,2,3) is the label of various algorithm for image enhancement, and the quantity of algorithm for image enhancement is equal to 3.
Step 4: it successively calculates in the case where each enhances algorithm, αijAverage value UjWith variance Sj:
Histogram equalization E1:
Contrast-limited histogram equalization E2:
Unsupervised color model E3
Step 5: the coefficient of variation of the enhancer image set under each algorithm for image enhancement is calculated separately out:
Histogram equalization E1:
Contrast-limited histogram equalization E2:
Unsupervised color model E3:
Step 6: it obtains after the coefficient of variation of enhancer image set, picking out most under each algorithm for image enhancement
The small coefficient of variation calculates, COV by MATLAB program1=0.0550, COV2=0.0539, COV3=0.0820.It follows that
COV2Minimum, with COV2Corresponding contrast-limited histogram equalization algorithm is the optimal enhancing algorithm of subgraph image set B.
Step 7: the evaluation and test system guided according to subset, contrast-limited histogram equalization algorithm are similarly original image
Collect the optimal enhancing algorithm of A.
In conclusion the present invention can provide one kind preferably for a variety of quality enhancement algorithms that can be applied to original image set
Strategy, to pick out a kind of optimal algorithm for image enhancement for large-scale image set.
Claims (1)
1. a kind of enhance preferred method based on subset guiding and the image set of the coefficient of variation, which comprises the following steps:
Step 1: grab sample is carried out to original image set A and obtains subgraph image set B, original image set A includes p width image, subgraph
Image set B includes n width image；
Step 2: a variety of algorithm for image enhancement (E are used1, E2..., Em) sub- image set B is enhanced, obtain multiple enhancings
Subgraph image set (B afterwards1, B2..., Bm), m is the quantity of algorithm for image enhancement；
Step 3: using a certain kind image quality evaluation criterion Q successively to enhanced subgraph image set (B1, B2..., Bm) carry out
Marking, marking result are denoted as αij, wherein i be image label, i=1,2 ... n, j be various algorithm for image enhancement label, j=
1,2 ... m；
Step 4: successively calculating enhances algorithm (E at each1, E2..., Em) under, αijAverage value UjWith variance Sj:
Step 5: it calculates in each algorithm for image enhancement (E1, E2..., Em) under enhancer image set the coefficient of variation, In
The coefficient of variation of enhancer image set is defined as COV under jth kind algorithm for image enhancementj, then have:
Step 6: when obtaining each algorithm for image enhancement (E1, E2..., Em) under enhancer image set the coefficient of variation after,
Pick out minimum coefficient of variation COVmin=min (COV1, COV2..., COVm), minimum coefficient of variation COVminCorresponding image
Enhancing algorithm is the optimal enhancing algorithm of the subgraph image set；
Step 7: after obtaining the optimal enhancing algorithm of subgraph image set B, according to subset guide evaluation and test system, subgraph image set B's
Optimal enhancing algorithm is the optimal enhancing algorithm of original image set A.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910594023.2A CN110415217A (en) | 2019-07-03 | 2019-07-03 | Enhance preferred method based on subset guiding and the image set of the coefficient of variation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910594023.2A CN110415217A (en) | 2019-07-03 | 2019-07-03 | Enhance preferred method based on subset guiding and the image set of the coefficient of variation |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110415217A true CN110415217A (en) | 2019-11-05 |
Family
ID=68358714
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910594023.2A Pending CN110415217A (en) | 2019-07-03 | 2019-07-03 | Enhance preferred method based on subset guiding and the image set of the coefficient of variation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110415217A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111179238A (en) * | 2019-12-24 | 2020-05-19 | 东华大学 | Subset confidence ratio dynamic selection method for subset-oriented guidance consistency enhancement evaluation |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7840066B1 (en) * | 2005-11-15 | 2010-11-23 | University Of Tennessee Research Foundation | Method of enhancing a digital image by gray-level grouping |
| CN103679157A (en) * | 2013-12-31 | 2014-03-26 | 电子科技大学 | Human face image illumination processing method based on retina model |
| CN108416755A (en) * | 2018-03-20 | 2018-08-17 | 南昌航空大学 | A kind of image de-noising method and system based on deep learning |
| CN109544487A (en) * | 2018-09-30 | 2019-03-29 | 西安电子科技大学 | A kind of infrared image enhancing method based on convolutional neural networks |
| CN109859180A (en) * | 2019-01-25 | 2019-06-07 | 东华大学 | Merge the image set quality enhancing evaluation method of a variety of measurement criterions |
-
2019
- 2019-07-03 CN CN201910594023.2A patent/CN110415217A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7840066B1 (en) * | 2005-11-15 | 2010-11-23 | University Of Tennessee Research Foundation | Method of enhancing a digital image by gray-level grouping |
| CN103679157A (en) * | 2013-12-31 | 2014-03-26 | 电子科技大学 | Human face image illumination processing method based on retina model |
| CN108416755A (en) * | 2018-03-20 | 2018-08-17 | 南昌航空大学 | A kind of image de-noising method and system based on deep learning |
| CN109544487A (en) * | 2018-09-30 | 2019-03-29 | 西安电子科技大学 | A kind of infrared image enhancing method based on convolutional neural networks |
| CN109859180A (en) * | 2019-01-25 | 2019-06-07 | 东华大学 | Merge the image set quality enhancing evaluation method of a variety of measurement criterions |
Non-Patent Citations (1)
| Title |
|---|
| EKIN DC ET AL: "《AutoAugment:Learning Augmentation Strategies from Data》", 《ARXIV》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111179238A (en) * | 2019-12-24 | 2020-05-19 | 东华大学 | Subset confidence ratio dynamic selection method for subset-oriented guidance consistency enhancement evaluation |
| CN111179238B (en) * | 2019-12-24 | 2022-12-20 | 东华大学 | Subset confidence ratio dynamic selection method for underwater image set-oriented guidance consistency enhancement evaluation |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108229550B (en) | Cloud picture classification method based on multi-granularity cascade forest network | |
| CN112215822B (en) | Face image quality evaluation method based on lightweight regression network | |
| CN109389180A (en) | A power equipment image-recognizing method and inspection robot based on deep learning | |
| CN110210294A (en) | Evaluation method, device, storage medium and the computer equipment of Optimized model | |
| CN111310756A (en) | Damaged corn particle detection and classification method based on deep learning | |
| CN111027377B (en) | Double-flow neural network time sequence action positioning method | |
| CN111126470B (en) | Image data iterative cluster analysis method based on depth measurement learning | |
| CN108427713A (en) | A kind of video summarization method and system for homemade video | |
| CN110827312A (en) | Learning method based on cooperative visual attention neural network | |
| CN112819015A (en) | Image quality evaluation method based on feature fusion | |
| CN114120094A (en) | Water pollution identification method and system based on artificial intelligence | |
| CN115393664A (en) | Active learning sample selection method for target detection | |
| CN115908421A (en) | Active learning medical image segmentation method based on superpixels and diversity | |
| CN110415217A (en) | Enhance preferred method based on subset guiding and the image set of the coefficient of variation | |
| CN104484679A (en) | Non-standard gun shooting bullet trace image automatic identification method | |
| CN109859180A (en) | Merge the image set quality enhancing evaluation method of a variety of measurement criterions | |
| CN115830514B (en) | Whole river reach surface flow velocity calculation method and system suitable for curved river channel | |
| CN116883718A (en) | Sugarcane seedling missing detection positioning method based on improved YOLOV7 | |
| CN116385935A (en) | Abnormal event detection algorithm based on unsupervised domain self-adaption | |
| CN115063679B (en) | Pavement quality assessment method based on deep learning | |
| CN116563205A (en) | Wheat spike counting detection method based on small target detection and improved YOLOv5 | |
| CN113869463B (en) | Long tail noise learning method based on cross enhancement matching | |
| CN111813996B (en) | Video searching method based on sampling parallelism of single frame and continuous multi-frame | |
| CN109376619A (en) | A kind of cell detection method | |
| CN112183752B (en) | End-to-end multi-example learning method based on automatic example selection |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20191105 |
|
| RJ01 | Rejection of invention patent application after publication |