CN114359244A - Image significance detection method based on super-pixel segmentation and multiple color features - Google Patents
Image significance detection method based on super-pixel segmentation and multiple color features Download PDFInfo
- Publication number
- CN114359244A CN114359244A CN202210020811.2A CN202210020811A CN114359244A CN 114359244 A CN114359244 A CN 114359244A CN 202210020811 A CN202210020811 A CN 202210020811A CN 114359244 A CN114359244 A CN 114359244A
- Authority
- CN
- China
- Prior art keywords
- super
- pixel
- color
- image
- time 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
- 238000001514 detection method Methods 0.000 title claims abstract description 21
- 230000011218 segmentation Effects 0.000 title claims abstract description 17
- 238000000034 method Methods 0.000 claims abstract description 20
- 238000012545 processing Methods 0.000 claims abstract description 10
- 230000004927 fusion Effects 0.000 claims abstract description 7
- 238000004364 calculation method Methods 0.000 claims description 13
- 239000003086 colorant Substances 0.000 claims description 9
- 230000000694 effects Effects 0.000 claims description 6
- 230000000007 visual effect Effects 0.000 description 5
- 238000012549 training Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000007261 regionalization Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Images
Landscapes
- Image Analysis (AREA)
Abstract
The invention discloses an image significance detection method based on superpixel segmentation and multiple color features. The method comprises the steps of performing superpixel segmentation on a real-time image by using a SLIC method to obtain a superpixel set and a corresponding label graph; processing the super-pixel set based on the label graph to obtain the pixel set, color characteristics and position information of each super-pixel in the super-pixel set; then calculating the contrast significant value of each super pixel, and performing gray assignment on the corresponding super pixel according to the contrast significant value of each super pixel to obtain a significant image; then, processing edge pixels of the current real-time image by using a main color description method to obtain an image edge main color, and then calculating to obtain a background prior image; and finally, performing linear fusion on the saliency map and the background prior map to obtain a final saliency map of the current real-time image. The invention improves the overall efficiency of the algorithm and the significance detection accuracy in the complex environment.
Description
Technical Field
The invention belongs to a method for detecting the saliency of an image in the field of computer vision, and particularly relates to a method for detecting the saliency of an image based on superpixel segmentation and multiple color characteristics.
Background
The image is used as an important information carrier, the understanding and processing of the image are always the key points of computer vision research, and the saliency detection of the image is one of the key points of the analysis of the image processed by computer vision. The saliency detection means that a visual attention mechanism of a human visual system is simulated through a model and an algorithm, and a region which is more likely to be an interested target is found out from an image, so that limited computing resources are more distributed to the region, the computing efficiency is improved, and the result of image processing is closer to the recognition effect of human vision.
Based on the visual attention mechanism, the visual saliency detection algorithm is mainly divided into a bottom-up algorithm and a top-up algorithm. The bottom-up method is mainly based on the bottom visual features and the contrast of images, the top-down method is mainly realized through machine learning or deep learning, and compared with the bottom-up method, the method needs manual labeling diagrams and a large number of test sets for training, and although the two saliency detection technologies can well realize saliency detection, certain problems still exist: the bottom-up method mainly depends on image color and contrast, and has poor detection effect in a complex scene with similar background and significant target color; the top-down method requires a lot of time for model training, and requires high computational power for the device during detection.
Disclosure of Invention
In order to solve the problems and needs in the background art, the present invention provides an image saliency detection method based on superpixel segmentation and multiple color features.
In order to achieve the purpose, the technical scheme of the invention comprises the following specific contents:
the invention comprises the following steps:
the first step is as follows: performing superpixel segmentation on an input real-time image through a Simple Linear Iterative Clustering (SLIC) algorithm to obtain a superpixel set and a corresponding label graph of the current real-time image;
the second step is that: processing a super-pixel set of the current real-time image based on the label graph to obtain a pixel set, color characteristics and position information of each super-pixel in the super-pixel set of the current real-time image; the color characteristics of each super pixel of the current real-time image are the average value of each color component of each super pixel in a CIELab color space and the average color in a quantized HSV color space;
the third step: calculating a significant value of each super pixel based on the contrast based on the pixel set of each super pixel of the current real-time image, the average value of each color component in the CIELab color space and the position information, and then calculating and obtaining a significant image of the current real-time image according to the significant value of each super pixel based on the contrast;
the fourth step: based on the average color of each super pixel of the current real-time image in a quantized HSV color space, processing the edge pixel of the current real-time image by using a main color description method based on the quantized HSV color space to obtain the image edge main color of the current real-time image, and calculating according to the image edge main color of the current real-time image to obtain a background prior image of the current real-time image;
the fifth step: respectively normalizing the saliency map and the background prior map of the current real-time image and then performing linear fusion to obtain the final saliency map of the current real-time image.
The third step is specifically as follows:
3.1) the position information of each super pixel is the geometric center of each super pixel; calculating the contrast between each super pixel and each super pixel according to the pixel set of each super pixel, the average value of each color component of each super pixel in the CIELab color space and the geometric center of each super pixel;
3.2) calculating a significant value based on the contrast of the current super pixel according to the contrast of the current super pixel;
3.3) repeating the step 3.2), traversing the residual superpixels, and calculating and obtaining the significant values of the residual superpixels based on the contrast;
and 3.4) carrying out gray assignment on the corresponding super-pixels according to the significant value of each super-pixel based on the contrast, and then obtaining a significant image of the current real-time image.
The fourth step is specifically as follows:
4.1) including edge pixels p of the real-time image in the superpixel seteAs an edge superpixel set E; calculating a color histogram of the edge superpixel set E based on the quantized HSV color space;
4.2) edge color L quantized in the color histogram of the edge superpixel set EESatisfy LE∈[0,71]The current quantized edge color LENeighborhood color of { L }E-1,LE,LE+1 the sum of the percentages in the color histogram of the edge superpixel set E is taken as the current quantized edge color LEThe percentage of the color histogram in the edge super-pixel set E;
4.3) repeating step 4.2), and traversing and calculating all quantized edge colors LEThe percentage of the color histogram in the edge super-pixel set E;
4.4) taking the percentage of the color histogram of the edge super-pixel set E which is more than or equal to 20 percent in the quantized edge colors as the image edge main color of the current real-time image;
4.5) calculating the significant value of each super pixel relative to the background information by combining the image edge main color of the current real-time image based on the average color in the quantized HSV color space in the color features of each super pixel of the current real-time image;
and 4.6) carrying out gray assignment on the corresponding super pixels based on the significant value of each super pixel relative to the background information to obtain a background prior image of the current real-time image.
In the fifth step, the calculation formula of linear fusion is as follows:
Sal(Sk)＝λSal1Sal'E(Sk)+λSal2Sal'C(Sk)
wherein Sal (S)k) Superpixel S in final saliency map representing current real-time imagekOf Sal'E(Sk) Normalized superpixel S in saliency map representing current real-time imagekSignificant value relative to background information, Sal'E(Sk) Super-pixel S in saliency map representing current real-time imagekNormalized significant value, Sal 'against background information'C(Sk) Superpixel S in background prior image representing current real-time imagekIs based on a normalized saliency value, lambda, of the contrastSal1Is a first weight coefficient, λSal2Is a second weight coefficient and satisfies lambdaSal1+λSal2＝1。
The calculation formula of the significant value of the current super pixel based on the contrast is as follows:
wherein, SalC(Sk) Representing a super-pixel SkK represents the total number of superpixels, w (S)k,Si) Representing a super-pixel SkAnd super pixel SiWeight coefficient between dpos(Sk,Si) Representing a super-pixel SkAnd super pixel SiCoefficient of spatial distance between, dc(Sk,Si) Representing a super-pixel SkAnd super pixel SiThe color contrast in CIELab color space, # { } represents the number of elements in the set, λposIs the coefficient that adjusts the effect of spatial distance on the saliency value calculation,
respectively representing super-pixels SkThe luminance in the CIELab color space and the average values of the first and second color channels,
respectively representing super-pixels SiThe luminance in the CIELab color space and the average values of the first and second color channels,
representing a super-pixel SkThe geometric center of (a) in the real-time image,
representing a super-pixel SiThe geometric center of (a) in the real-time image, λposIs the coefficient that adjusts the significant value of the spatial distance that affects the contrast, and i is the superpixel number in the superpixel set.
The calculation formula of the significant value of each super pixel relative to the background information is as follows:
wherein, SalE(Sk) Representing a super-pixel SkThe significance value with respect to the background information,
representing a super-pixel SkThe average color in the quantized HSV color space,
the method comprises the steps of quantizing the ith quantized edge color in the main color of the edge of the image, | | represents absolute value operation, min represents minimum value operation, and N is the color number of the main color of the edge of the image.
Compared with the prior art, the invention has the following beneficial effects:
1. the detection effect in a complex environment with similar background and target colors is improved by a saliency detection method based on super-pixel segmentation and multiple color space features.
2. The image area significant value is calculated based on the super-pixel segmentation regionalization description image information, the pixel set of each super-pixel, the average value of each color component of each super-pixel in the CIELab color space, the average color in the quantized HSV color space and the geometric center of each super-pixel, and the overall calculation efficiency and accuracy are improved.
3. And the HSV color space with better effect in color segmentation is used for describing background prior information, so that the significance detection accuracy in a complex environment is improved.
Drawings
FIG. 1 is a flow chart of a saliency detection algorithm based on superpixel segmentation and multiple color features of the present invention.
Fig. 2 is a saliency detection raw image of the present invention.
Fig. 3 is a saliency detection result image of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention. In addition, the technical features involved in the embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.
The invention will now be further described with reference to the following examples and drawings:
the implementation technical scheme of the invention is as follows:
as shown in fig. 1, the present invention comprises the steps of:
the first step is as follows: performing superpixel segmentation on an input real-time image through a Simple Linear Iterative Clustering (SLIC) algorithm to obtain a superpixel set and a corresponding label graph of the current real-time image; the real-time image is decomposed into a plurality of local areas, the calculation amount of subsequent calculation is reduced, and the overall efficiency of the algorithm is improved. The input real-time image is shown in fig. 2.
The second step is that: processing a super-pixel set of the current real-time image based on the label graph to obtain a pixel set, color characteristics and position information of each super-pixel in the super-pixel set of the current real-time image; the color characteristics of each super pixel of the current real-time image are the average value of each color component of each super pixel in a CIELab color space and the average color in a quantized HSV color space;
the third step: calculating a significant value of each super pixel based on the contrast based on the pixel set of each super pixel of the current real-time image, the average value of each color component in the CIELab color space and the position information, and then calculating and obtaining a significant image of the current real-time image according to the significant value of each super pixel based on the contrast;
the third step is specifically as follows:
3.1) the position information of each super pixel is the geometric center of each super pixel; calculating the contrast between each super pixel and each super pixel according to the pixel set of each super pixel, the average value of each color component of each super pixel in the CIELab color space and the geometric center of each super pixel;
3.2) calculating the significant value of the current superpixel based on the contrast of the current superpixel by the following formula:
wherein, SalC(Sk) Representing a super-pixel SkK represents the total number of super pixels, and in this embodiment, K is 150. w (S)k,Si) Representing a super-pixel SkAnd super pixel SiWeight coefficient between dpos(Sk,Si) Representing a super-pixel SkAnd super pixel SiCoefficient of spatial distance between, dc(Sk,Si) Representing a super-pixel SkAnd super pixel SiThe color contrast in CIELab color space, # { } represents the number of elements in the set, λposIs the coefficient that adjusts the influence of the spatial distance on the calculation of the saliency value, in this example λpos＝10。
Respectively representing super-pixels SkThe luminance in the CIELab color space and the average values of the first and second color channels,
respectively representing super-pixels SiThe luminance in the CIELab color space and the average of the first and second color channels, the first color channel being from dark green (low luminance value) to gray (medium luminance value) to bright pink red (high luminance value)Value), the second color channel from bright blue (low luminance value) to gray (medium luminance value) to yellow (high luminance value),
representing a super-pixel SkThe geometric center of (a) in the real-time image,
representing a super-pixel SiThe geometric center of (a) in the real-time image, λposThe coefficient is a significant value of adjusting the spatial distance to influence the contrast, and i is a super pixel serial number in the super pixel set;
3.3) repeating the step 3.2), traversing the residual superpixels, and calculating and obtaining the significant values of the residual superpixels based on the contrast;
and 3.4) carrying out gray assignment on the corresponding super-pixels according to the significant value of each super-pixel based on the contrast, and then obtaining a significant image of the current real-time image.
The fourth step: based on the average color of each super pixel of the current real-time image in a quantized HSV color space, processing the edge pixel of the current real-time image by using a main color description method based on the quantized HSV color space to obtain the image edge main color of the current real-time image, and describing the background information of the real-time image by using the image edge main color; calculating to obtain a background prior image of the current real-time image according to the image edge main color of the current real-time image;
the fourth step is specifically as follows:
4.1) including edge pixels p of the real-time image in the superpixel seteAs an edge superpixel set E; satisfy the requirement of
SjRepresenting the jth superpixel, P, in the edge superpixel set EjRepresenting the pixel of the jth superpixel in the edge superpixel set E. Calculating a color histogram of the edge superpixel set E based on the quantized HSV color space;
4.2) at the edgeQuantized edge color L in color histogram of super-pixel set EESatisfy LE∈[0,71]The current quantized edge color LENeighborhood color of { L }E-1,LE,LE+1 the sum of the percentages in the color histogram of the edge superpixel set E is taken as the current quantized edge color LEThe percentage of the color histogram in the edge super-pixel set E;
4.3) repeating step 4.2), and traversing and calculating all quantized edge colors LEThe percentage of the color histogram in the edge super-pixel set E;
4.4) taking the percentage of the color histogram of the edge super-pixel set E in the quantized edge colors to be more than or equal to 20 percent as the image edge main color D of the current real-time imageEDCDescribing the background information of the real-time image by using the edge dominant color of the image;
wherein N is the number of colors of the dominant color of the edge of the image and satisfies N ∈ {0,1,2,3,4,5},
representing an nth quantized edge color of the edge dominant colors of the image;
4.5) calculating the significant value of each super pixel relative to the background information by combining the image edge main color of the current real-time image based on the average color in the quantized HSV color space in the color features of each super pixel of the current real-time image through the following formula:
wherein, SalE(Sk) Representing a super-pixel SkThe significance value with respect to the background information,
representing a super-pixel SkThe average color in the quantized HSV color space,
the method comprises the steps that the ith quantized edge color in the main color of the edge of the image is obtained, | | represents absolute value operation, min represents minimum value operation, and N is the color number of the main color of the edge of the image;
and 4.6) carrying out gray assignment on the corresponding super pixels based on the significant value of each super pixel relative to the background information to obtain a background prior image of the current real-time image.
The fifth step: the saliency map and the background prior map of the current real-time image are normalized respectively and then subjected to linear fusion to obtain a final saliency map of the current real-time image, as shown in fig. 3.
In the fifth step, the calculation formula of linear fusion is as follows:
Sal(Sk)＝λSal1Sal'E(Sk)+λSal2Sal'C(Sk)
wherein Sal (S)k) Superpixel S in final saliency map representing current real-time imagekOf Sal'E(Sk) Super-pixel S in saliency map representing current real-time imagekNormalized significant value, Sal 'against background information'C(Sk) Superpixel S in background prior image representing current real-time imagekIs based on a normalized saliency value, lambda, of the contrastSal1Is a first weight coefficient, λSal2Is a second weight coefficient and satisfies lambdaSal1+λSal2＝1。
The above-mentioned contents are only technical ideas of the present invention, and the protection scope of the present invention is not limited thereby, and any modifications made on the basis of the technical ideas proposed by the present invention fall within the protection scope of the claims of the present invention.
Claims (6)
1. An image saliency detection method based on superpixel segmentation and multiple color features is characterized by comprising the following steps:
the first step is as follows: performing superpixel segmentation on an input real-time image through a simple linear iterative clustering algorithm to obtain a superpixel set and a corresponding label graph of the current real-time image;
the second step is that: processing a super-pixel set of the current real-time image based on the label graph to obtain a pixel set, color characteristics and position information of each super-pixel in the super-pixel set of the current real-time image; the color characteristics of each super pixel of the current real-time image are the average value of each color component of each super pixel in a CIELab color space and the average color in a quantized HSV color space;
the third step: calculating a significant value of each super pixel based on the contrast based on the pixel set of each super pixel of the current real-time image, the average value of each color component in the CIELab color space and the position information, and then calculating and obtaining a significant image of the current real-time image according to the significant value of each super pixel based on the contrast;
the fourth step: based on the average color of each super pixel of the current real-time image in a quantized HSV color space, processing the edge pixel of the current real-time image by using a main color description method based on the quantized HSV color space to obtain the image edge main color of the current real-time image, and calculating according to the image edge main color of the current real-time image to obtain a background prior image of the current real-time image;
the fifth step: respectively normalizing the saliency map and the background prior map of the current real-time image and then performing linear fusion to obtain the final saliency map of the current real-time image.
2. The image saliency detection method based on superpixel segmentation and multiple color features according to claim 1, characterized in that said third step specifically is:
3.1) the position information of each super pixel is the geometric center of each super pixel; calculating the contrast between each super pixel and each super pixel according to the pixel set of each super pixel, the average value of each color component of each super pixel in the CIELab color space and the geometric center of each super pixel;
3.2) calculating a significant value based on the contrast of the current super pixel according to the contrast of the current super pixel;
3.3) repeating the step 3.2), traversing the residual superpixels, and calculating and obtaining the significant values of the residual superpixels based on the contrast;
and 3.4) carrying out gray assignment on the corresponding super-pixels according to the significant value of each super-pixel based on the contrast, and then obtaining a significant image of the current real-time image.
3. The image saliency detection method based on superpixel segmentation and multiple color features according to claim 1, characterized in that said fourth step specifically is:
4.1) including edge pixels p of the real-time image in the superpixel seteAs an edge superpixel set E; calculating a color histogram of the edge superpixel set E based on the quantized HSV color space;
4.2) edge color L quantized in the color histogram of the edge superpixel set EESatisfy LE∈[0,71]The current quantized edge color LENeighborhood color of { L }E-1,LE,LE+1 the sum of the percentages in the color histogram of the edge superpixel set E is taken as the current quantized edge color LEThe percentage of the color histogram in the edge super-pixel set E;
4.3) repeating step 4.2), and traversing and calculating all quantized edge colors LEThe percentage of the color histogram in the edge super-pixel set E;
4.4) taking the percentage of the color histogram of the edge super-pixel set E which is more than or equal to 20 percent in the quantized edge colors as the image edge main color of the current real-time image;
4.5) calculating the significant value of each super pixel relative to the background information by combining the image edge main color of the current real-time image based on the average color in the quantized HSV color space in the color features of each super pixel of the current real-time image;
and 4.6) carrying out gray assignment on the corresponding super pixels based on the significant value of each super pixel relative to the background information to obtain a background prior image of the current real-time image.
4. The method for detecting image saliency based on super-pixel segmentation and multiple color features according to claim 1, wherein in the fifth step, the calculation formula of linear fusion is as follows:
Sal(Sk)＝λSal1Sal'E(Sk)+λSal2Sal'C(Sk)
wherein Sal (S)k) Superpixel S in final saliency map representing current real-time imagekOf Sal'E(Sk) Normalized superpixel S in saliency map representing current real-time imagekSignificant value relative to background information, Sal'E(Sk) Super-pixel S in saliency map representing current real-time imagekNormalized significant value, Sal 'against background information'C(Sk) Superpixel S in background prior image representing current real-time imagekIs based on a normalized saliency value, lambda, of the contrastSal1Is a first weight coefficient, λSal2Is a second weight coefficient and satisfies lambdaSal1+λSal2＝1。
5. The method according to claim 2, wherein the calculation formula of the saliency value based on contrast of the current super pixel is as follows:
wherein, SalC(Sk) Representing a super-pixel SkK represents the total number of superpixels, w (S)k,Si) Representing a super-pixel SkAnd super pixel SiWeight coefficient between dpos(Sk,Si) Representing a super-pixel SkAnd super pixel SiCoefficient of spatial distance between, dc(Sk,Si) Representing a super-pixel SkAnd super pixel SiThe color contrast in CIELab color space, # { } represents the number of elements in the set, λposIs the coefficient that adjusts the effect of spatial distance on the saliency value calculation,
respectively representing super-pixels SkThe luminance in the CIELab color space and the average values of the first and second color channels,
respectively representing super-pixels SiThe luminance in the CIELab color space and the average values of the first and second color channels,
representing a super-pixel SkThe geometric center of (a) in the real-time image,
representing a super-pixel SiThe geometric center of (a) in the real-time image, λposIs a coefficient for adjusting the significance of the spatial distance influencing the contrastAnd i is the superpixel number in the superpixel set.
6. The method according to claim 3, wherein the calculation formula of the saliency value of each super pixel with respect to the background information is as follows:
wherein, SalE(Sk) Representing a super-pixel SkThe significance value with respect to the background information,
representing a super-pixel SkThe average color in the quantized HSV color space,
the method comprises the steps of quantizing the ith quantized edge color in the main color of the edge of the image, | | represents absolute value operation, min represents minimum value operation, and N is the color number of the main color of the edge of the image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210020811.2A CN114359244A (en) | 2022-01-10 | 2022-01-10 | Image significance detection method based on super-pixel segmentation and multiple color features |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210020811.2A CN114359244A (en) | 2022-01-10 | 2022-01-10 | Image significance detection method based on super-pixel segmentation and multiple color features |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114359244A true CN114359244A (en) | 2022-04-15 |
Family
ID=81107619
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210020811.2A Pending CN114359244A (en) | 2022-01-10 | 2022-01-10 | Image significance detection method based on super-pixel segmentation and multiple color features |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114359244A (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170083762A1 (en) * | 2015-06-22 | 2017-03-23 | Photomyne Ltd. | System and Method for Detecting Objects in an Image |
| CN107025672A (en) * | 2017-03-30 | 2017-08-08 | 上海理工大学 | A kind of conspicuousness detection method based on improvement convex closure |
| CN107578451A (en) * | 2017-09-20 | 2018-01-12 | 太原工业学院 | A kind of adaptive key color extraction method towards natural image |
| CN108549891A (en) * | 2018-03-23 | 2018-09-18 | 河海大学 | Multi-scale diffusion well-marked target detection method based on background Yu target priori |
| CN109242854A (en) * | 2018-07-14 | 2019-01-18 | 西北工业大学 | A kind of image significance detection method based on FLIC super-pixel segmentation |
| CN112037230A (en) * | 2019-06-04 | 2020-12-04 | 北京林业大学 | Forest region image segmentation algorithm based on super-pixel and super-metric contour map |
| CN113705579A (en) * | 2021-08-27 | 2021-11-26 | 河海大学 | Automatic image annotation method driven by visual saliency |
-
2022
- 2022-01-10 CN CN202210020811.2A patent/CN114359244A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170083762A1 (en) * | 2015-06-22 | 2017-03-23 | Photomyne Ltd. | System and Method for Detecting Objects in an Image |
| CN107025672A (en) * | 2017-03-30 | 2017-08-08 | 上海理工大学 | A kind of conspicuousness detection method based on improvement convex closure |
| CN107578451A (en) * | 2017-09-20 | 2018-01-12 | 太原工业学院 | A kind of adaptive key color extraction method towards natural image |
| CN108549891A (en) * | 2018-03-23 | 2018-09-18 | 河海大学 | Multi-scale diffusion well-marked target detection method based on background Yu target priori |
| CN109242854A (en) * | 2018-07-14 | 2019-01-18 | 西北工业大学 | A kind of image significance detection method based on FLIC super-pixel segmentation |
| CN112037230A (en) * | 2019-06-04 | 2020-12-04 | 北京林业大学 | Forest region image segmentation algorithm based on super-pixel and super-metric contour map |
| CN113705579A (en) * | 2021-08-27 | 2021-11-26 | 河海大学 | Automatic image annotation method driven by visual saliency |
Non-Patent Citations (6)
| Title |
|---|
| JIWHAN KIM等: "Salient Region Detection via High-Dimensional Color Transform", 《 IEEE TRANSACTIONS ON IMAGE PROCESSING 》, 26 October 2015 (2015-10-26) * |
| JIWHAN KIM等: "Salient region detection via high-dimensional color transform", 《2014 IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION》, 25 September 2014 (2014-09-25) * |
| 司马海峰;米爱中;王志衡;杜守恒;: "显著特征融合的主颜色聚类分割算法", 模式识别与人工智能, no. 06, 15 June 2016 (2016-06-15) * |
| 孙赫赫;尚晓清;王冲;: "基于区域对比的图像显著性检测方法", 计算机工程与应用, no. 10, 19 April 2017 (2017-04-19) * |
| 肖春霞;聂勇伟;黄先锋;赵勇;彭群生;: "基于联合双边滤波的纹理合成上采样算法", 计算机学报, no. 02, 15 February 2009 (2009-02-15) * |
| 邵凯旋;余映;钱俊;吴青龙;杨鉴;: "基于边缘信息结合空间权重的图像显著性检测算法研究", 云南大学学报(自然科学版), no. 03, 10 May 2020 (2020-05-10) * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109829443B (en) | Video behavior identification method based on image enhancement and 3D convolution neural network | |
| WO2022199583A1 (en) | Image processing method and apparatus, computer device, and storage medium | |
| CN108629783B (en) | Image segmentation method, system and medium based on image feature density peak search | |
| CN114359323B (en) | Image target area detection method based on visual attention mechanism | |
| CN112614060A (en) | Method and device for rendering human face image hair, electronic equipment and medium | |
| CN111008632B (en) | License plate character segmentation method based on deep learning | |
| CN107292933B (en) | Vehicle color identification method based on BP neural network | |
| CN102306307B (en) | Positioning method of fixed point noise in color microscopic image sequence | |
| CN112435191A (en) | Low-illumination image enhancement method based on fusion of multiple neural network structures | |
| CN111583279A (en) | Super-pixel image segmentation method based on PCBA | |
| CN115131375B (en) | Automatic ore cutting method | |
| CN108711160B (en) | Target segmentation method based on HSI (high speed input/output) enhanced model | |
| CN112464731A (en) | Traffic sign detection and identification method based on image processing | |
| CN114821452B (en) | Colored drawing train number identification method, system and medium | |
| CN113947732B (en) | Aerial visual angle crowd counting method based on reinforcement learning image brightness adjustment | |
| CN107146258B (en) | Image salient region detection method | |
| CN110415816B (en) | Skin disease clinical image multi-classification method based on transfer learning | |
| CN109766860B (en) | Face detection method based on improved Adaboost algorithm | |
| CN114359244A (en) | Image significance detection method based on super-pixel segmentation and multiple color features | |
| CN111797694A (en) | License plate detection method and device | |
| CN109165659B (en) | Vehicle color identification method based on superpixel segmentation | |
| CN108154188B (en) | FCM-based artificial text extraction method under complex background | |
| CN115761459A (en) | Multi-scene self-adaption method for bridge and tunnel apparent disease identification | |
| CN115496816A (en) | Sequence clothing image theme color self-adaptive extraction method | |
| CN112487927B (en) | Method and system for realizing indoor scene recognition based on object associated attention |
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 | ||
| CB02 | Change of applicant information | ||
| CB02 | Change of applicant information |
Country or region after: China Address after: 316021 Zhoushan campus of Zhejiang University, No.1 Zheda Road, Dinghai District, Zhoushan City, Zhejiang Province Applicant after: ZHEJIANG University Address before: 310058 Yuhang Tang Road, Xihu District, Hangzhou, Zhejiang 866 Applicant before: ZHEJIANG University Country or region before: China |