CN110288617A - Based on the shared sliced image of human body automatic division method for scratching figure and ROI gradual change - Google Patents
Based on the shared sliced image of human body automatic division method for scratching figure and ROI gradual change Download PDFInfo
- Publication number
- CN110288617A CN110288617A CN201910600636.2A CN201910600636A CN110288617A CN 110288617 A CN110288617 A CN 110288617A CN 201910600636 A CN201910600636 A CN 201910600636A CN 110288617 A CN110288617 A CN 110288617A
- Authority
- CN
- China
- Prior art keywords
- pixel
- point
- region
- color
- area
- 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
Classifications
-
- 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
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- 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
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/90—Determination of colour characteristics
-
- 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
- G06T2207/10024—Color 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
- G06T2207/20092—Interactive image processing based on input by user
- G06T2207/20104—Interactive definition of region of interest [ROI]
-
- 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
- G06T2207/30196—Human being; Person
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of based on the shared sliced image of human body automatic division method for scratching figure and ROI gradual change, comprising the following steps: S1: is scribbled roughly to area-of-interest；S2: the interior zone A of area-of-interest is chosen；S3: covering area-of-interest and the borderline region for reading area-of-interest；S4: being defined as white portion for the foreground area of interior zone A, middle section is defined as grey parts, background area is defined as black portions, obtains three components of area-of-interest；S5: the black and white mask of area-of-interest is obtained using shared stingy nomography according to three component information, obtains the prospect skeleton of black and white mask with thinning algorithm；S6: using prospect skeleton as the prospect of next picture scribble, three components of next picture are obtained using flooding；S7: circulation S5 to S6 obtains the black and white mask of all picture area-of-interests until having handled all pictures.
Description
Technical field
The present invention relates to technical field of image segmentation more particularly to a kind of human body based on shared stingy figure and ROI gradual change to cut
Picture automatic division method.
Background technique
Image segmentation is intended to extract foreground elements in the picture by color and opacity, is to carry out medical image segmentation
One of main means.Although possessing many progress in terms of raising is scratched as the accuracy of technology in recent years, there are still one
A little FAQs, such as scratch drawing method using alpha in the prior art and carry out image zooming-out, but need to give all images
Three components.In alpha segmentation, each unknown pixel searches for the entire border circular areas centered on oneself, and shared scratch is schemed
Algorithm is circle to be divided into many sectors, and adjacent pixel searches for nonoverlapping fan-shaped region, improves treatment effeciency.But due to
Using multiple images that this method is for sequence of partitions, therefore error can be generated in cutting procedure, and build up, make point
It is very accurate for cutting result not.
Summary of the invention
According to problem of the existing technology, cut the invention discloses a kind of based on the shared human body for scratching figure and ROI gradual change
Picture automatic division method, concrete scheme specifically includes the following steps:
S1: reading the area-of-interest of image, is scribbled roughly to area-of-interest；
S2: choosing the interior zone A of area-of-interest, calculates interior zone A and the color difference range by scribble pixel, record
By scribble pixel with the color minus tolerance maximum value loDiff1 of interior zone A pixel and by scribble pixel and interior zone A pixel
Color principal-employment maximum value upDiff1；
S3: covering area-of-interest and the borderline region for reading area-of-interest, by the sum of interior zone A and borderline region
It is defined as region B, searches region B and the color difference range by scribble pixel, record is born by scribble pixel and the color of region B pixel
Poor maximum value loDiff2 and by the color principal-employment maximum value upDiff2 of scribble pixel and region B pixel；
S4: interior zone A is defined as the part of white portion, region B more than region A and is defined as grey parts, B and does not wrap
The region contained is defined as black portions, obtains whole three component of picture；
S5: the black and white mask of area-of-interest is obtained using shared stingy nomography according to three component information, with parallel thin
Change algorithm and obtains the prospect skeleton of black and white mask；
S6: using prospect skeleton as next picture prospect scribble, the loDiff1 according to obtained in S2, S3 and
UpDiff1, loDiff2 and upDiff2 data obtain three components of next picture using flooding；
S7: circulation S5 to S6 obtains the black and white mask of all picture area-of-interests until having handled all pictures.
It is specifically adopted according to three component information using the black and white mask that shared stingy nomography obtains area-of-interest in the S5
With such as under type:
S51: the maximum quantity parameter K of setting foreground pixel and background pixel extracts K road from each unknown pixel
Diameter extends unknown pixel region to the rectangular area K × K, wherein the start angle in path is pressed if the angle in each path is π/K
According to cyclically-varying；
S52: it records when encountering foreground pixel or background pixel in each path for the first time, record foreground pixel and unknown point
Between image space and color space distance and the image space between background pixel and unknown point and color space away from
From when edge of the path beyond image then stops searching foreground and background pixel；
S53: color card is converted by candidate point；Calculate the gap of sample point and unknown point color；Unknown point is calculated to arrive
The number that pixel is mutated on straight line path between the foreground point and background dot of sampling；Calculate sampling foreground point and background dot with
Physical distance between unknown point；Foreground point and background dot are subjected to the sampling picture that optimal combination is found in minimum optimum organization
Element；
S54: data processing is carried out to the sampled pixel of optimal combination and obtains the smallest a pair of of prospect background data pair, by this
Data are to being defined as optimum sampling point；
S55: for the unknown pixel point in three component grey areas, radius threshold is set, in the radius threshold
Each unknown pixel point is counted to obtain three minimum MP values, obtains the corresponding three unknown pixels point of three minimum MP values
Relevant colors data are weighted the relevant colors data and average treatment obtain data T, and data T is defined as most preferably
Sampled point；
S56: new foreground pixel, background pixel, transparency and confidence information are obtained according to optimum sampling point and obtain feeling emerging
The black and white mask in interesting region.
Three components of next picture are obtained specifically in the following way using flooding in S6:
S61: the location information and color value information of reading prospect skeleton all pixels, setting loDiff1 and upDiff1,
LoDiff2 and upDiff2 value；
S62: the pixel Q in prospect skeleton is chosen, is begun looking for from the neighbor pixel point of four neighborhoods of pixel Q:
If the color minus tolerance of current pixel point Q and its neighborhood territory pixel is less than loDiff1, and current pixel point Q and its neighborhood territory pixel
Color principal-employment is less than upDiff1, then is added to this pixel Q and has chosen in set, until there is no any in iterative process
Until pixel meets above-mentioned condition, gained region is denoted as region D1；
S63: the pixel W in prospect skeleton is chosen, is looked into since the neighbor pixel point of four neighborhoods of pixel W point
It looks for, if the color minus tolerance of current pixel point W and its neighborhood territory pixel is less than loDiff2, and current pixel point W and its neighborhood territory pixel
Color principal-employment be less than upDiff2, then this pixel W is added to and has been chosen in set, until in iterative process there is no appoint
Until what pixel meets above-mentioned condition, gained region is denoted as region D2；
S64: region D1 is defined as the part of white portion in three components, region D2 greater than region D1 and is defined as grey portion
Point, rest part is defined as black portions.
By adopting the above-described technical solution, a kind of human body based on shared stingy figure and ROI gradual change provided by the invention is cut
Picture automatic division method determines the foreground part of next picture using skeleton, compared with seed point methods, the application
The lack part of disclosed method image is less, and divides fineness height；Area-of-interest is obtained using shared stingy nomography
Black and white mask, to reduce the calculation amount of partitioning algorithm, reduce resource consumption, improve splitting speed；Improve figure
The degree of automation of piece segmentation, since human intervention is less in image segmentation process, it is more flexible to operate, extraction accuracy is obtained
It improves, therefore can accurately and quickly extract the image of human organ, be subsequent organ three-dimensional modeling, and then be put to
Clinical application provides strong technical support.
Detailed description of the invention
In order to illustrate the technical solutions in the embodiments of the present application or in the prior art more clearly, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
The some embodiments recorded in application, for those of ordinary skill in the art, without creative efforts,
It is also possible to obtain other drawings based on these drawings.
Fig. 1 is the flow chart of the method for the present invention；
Fig. 2 is the schematic diagram of present invention image to be split；
Fig. 3 is the skeleton that the Freehandhand-drawing of foreground image of the present invention marks；
Fig. 4 is a part of region inside area-of-interest of the present invention；
Fig. 5 is the region that area-of-interest is completely covered in the present invention；
Fig. 6 is three components that the present invention is used for segmentation approximate region；
Fig. 7 is the schematic diagram in the present invention setting path k；
Fig. 8 is the black and white mask that the present invention once divides obtained area-of-interest；
Fig. 9 is the prospect skeleton that thinning algorithm of the present invention obtains；
Figure 10 is the picture to be split that the present invention gradually changes；
Figure 11 is the prospect black and white mask that the present invention gradually changes.
Specific embodiment
To keep technical solution of the present invention and advantage clearer, with reference to the attached drawing in the embodiment of the present invention, to this
Technical solution in inventive embodiments carries out clear and complete description:
A kind of sliced image of human body automatic division method based on shared stingy figure and ROI gradual change as shown in Figure 1 specifically wraps
Include following steps:
S1: reading the area-of-interest of image, carries out rough scribble to area-of-interest and obtains skeleton part, wherein first
Image to be split is opened as shown in Fig. 2, Freehandhand-drawing scribble part is as shown in Figure 3.
S2: it is appropriate to choose area-of-interest interior zone A, as shown in figure 4, the region is enable to include most of interested
Region.Find out the color difference range in the region and pixel of scribbling, the color minus tolerance maximum value of record scribble pixel and this area pixel
LoDiff1, and the color principal-employment maximum value upDiff1 of scribble pixel and this area pixel；
S3: covering area-of-interest and the borderline region for reading area-of-interest, by the sum of interior zone A and borderline region
It is defined as region B, searches region B and the color difference range by scribble pixel, record is born by scribble pixel and the color of region B pixel
Poor maximum value loDiff2 and by the color principal-employment maximum value upDiff2 of scribble pixel and region B pixel；
S4: interior zone A is defined as the part of white portion, region B more than region A and is defined as grey parts, B and does not wrap
The region contained is defined as black portions, obtains whole three component of picture；Three are generated according to the prospect approximate range that S2, S3 are determined
Component.Wherein S2 selected areas is three component white portions, and it is grey in three components that S3 selected portion, which subtracts S2 selected portion,
Color part, rest part are three component black portions.Three components are obtained, three components are as shown in Figure 6.
S5: the black and white mask of area-of-interest is obtained using shared stingy nomography according to three component information, is calculated with refinement
Method obtains the prospect skeleton of black and white mask；And skeleton is suitably trimmed.
S6: using prospect skeleton as next picture prospect scribble, the loDiff1 according to obtained in S2, S3 and
UpDiff1, loDiff2 and upDiff2 data obtain three components of next picture using flooding；
S7: circulation S5 to S6 obtains the black and white mask of all picture area-of-interests until having handled all pictures.Gradually
The picture to be split of variation and the prospect black and white mask gradually changed are as shown in Figure 10 and Figure 11.
Further, described to be had according to three component information using the black and white mask that shared stingy nomography obtains area-of-interest
Body is in the following way: for having been subjected to pretreated serializing image above, the region in S2 being utilized from pixel and is recycled to
External mode is extended on a small scale, to reduce the quantity of zone of ignorance, and reduces the calculation amount in later period
S51: the maximum quantity parameter K of setting foreground pixel and background pixel extracts K road from each unknown pixel
Diameter extends unknown pixel region to the rectangular area K × K, wherein the start angle in path is pressed if the angle in each path is π/K
According to cyclically-varying, as shown in Figure 7；
S52: it records when encountering foreground pixel or background pixel in each path for the first time, record foreground pixel and unknown point
Between image space and color space distance and the image space between background pixel and unknown point and color space away from
From when edge of the path beyond image then stops searching foreground and background pixel；
S53: the gap for converting color card for candidate point, calculating sample point and unknown point color calculates unknown point and arrives
The number that pixel is mutated on straight line path between the foreground point and background dot of sampling；Calculate sampling foreground point and background dot with
Physical distance between unknown point；Foreground point and background dot are subjected to the sampling picture that optimal combination is found in minimum optimum organization
Element.
The specific calculating process of S52 and S53 is: converting color card for candidate point by formula (1)；Pass through formula
(2) gap of sample point and unknown point color is calculated；By formula (3) calculate unknown point to sample foreground and background point it
Between straight line path on, pixel mutation number；By formula (4) calculate unknown point to sampling foreground point or background dot and
Physical distance between unknown point；Optimum organization is minimized by formula (5), finds the sampled pixel of optimal combination.
In view of the unknown point of straight line path between sample foreground point and background dot, to avoid pixel as few as possible prominent
Become, simplified using formula:
Select the pixel color along gained path after integrated.
Minimize optimum organization simultaneously:
In conjunction with this four conditions, following objective function is finally obtained:
The data sampled before processing, and one group of combination for minimizing target function value is recorded, and initially determine that best
Sampled point: eN=3, eA=2, ef=1, eb=4
S54: data processing is carried out to the sampled pixel of optimal combination and obtains the smallest a pair of of prospect background data pair, by this
Data are to being defined as optimum sampling point；
S55: for the unknown pixel point in three component grey areas, being arranged radius threshold, will be in the radius threshold
Each unknown pixel point using formula (2) statistics obtain three minimum MP values, obtain three minimum MP values it is corresponding three it is unknown
The relevant colors data of pixel are weighted the relevant colors data and average treatment obtain data T, and data T is determined
Justice is optimum sampling point；
Wherein weighted sum average treatment formula are as follows:
S56: new foreground pixel, background pixel, transparency and confidence information are obtained according to optimum sampling point and obtain feeling emerging
The black and white mask in interesting region.The black and white mask for obtaining picture area-of-interest carries out local smoothing method by Gaussian Blur, to reduce
Noise.Obtained black and white mask is as shown in Figure 8.
Further, the black and white binary result image obtained for shared stingy nomography, obtains prospect bone with thinning algorithm
Frame, specifically in the following way:
P1: for each foreground pixel, its eight neighborhood is numbered in certain sequence, which is set as P1, just on
Square pixel is P2, rest of pixels numbers increase in the direction of the clock.Extract skeleton thinning process be divided into two subprocess with
Guarantee the integrality of skeleton.
P2: thinning process P1, if the foreground pixel point P investigated1Neighborhood territory pixel meet following condition, then by its turn
At background pixel；
(a)2≤B(P1)≤6；(b)A(P1)=1；(c)P2×P4×P6=0；(d)P4×P6×P8=0
Wherein A (P1) it is { P in numerical order2,P3,…P9Number of the pixel by 0 change 1, B (P1) it is P1Non-zero neighbours
Number.
P3: further refinement: its conditional (a), it is (b) constant, (c), (d) it is changed to as follows:
(e)P2×P4×P8=0 (f) P2×P6×P8=0
If the foreground pixel investigated meets condition (a) (b) (e) (f), it is converted into background pixel.
P4: cycle P 1 arrives P3, and until there is no foreground pixels to be replaced by background pixel, thinning process terminates.It obtains
Prospect skeleton is as shown in Figure 9.
Further, prospect skeleton is scribbled as the prospect of next picture in S6, according to obtained in S2, S3
LoDiff1 and upDiff1, loDiff2 and upDiff2 data obtain three components, specific of next picture using flooding
In the following way:
S61: position and the color value of skeleton all pixels are obtained.Foreground pixel is set by pixel where all skeletons.
S62: setting loDiff1 and upDiff1, loDiff2 and upDiff2 value；
S63: the pixel Q in prospect skeleton is chosen, from the neighbor pixel point V of four neighborhoods for the pixel Q being selected
It begins looking for, if the color minus tolerance of current pixel point Q and its neighborhood territory pixel point V is less than loDiff1, and observes current pixel point
The color principal-employment of Q and its neighborhood territory pixel point V is less than upDiff1, it is added to and has chosen in set.
S64: so circulation, until having chosen set not to be further added by, i.e., until there is no any pictures in iterative process
Until element meets above-mentioned condition, gained region is denoted as D1.
S65: emptying foreground pixel, and the pixel in skeleton is added in foreground pixel again, from foreground pixel set
In selected pixels point W again, begun looking in the neighbor pixel point of four neighborhoods of pixel W, if current pixel point W and its neighborhood
The color minus tolerance of pixel is less than loDiff2, and searches current pixel point W and the color principal-employment of its neighborhood territory pixel point is less than
This point then is chosen to be added to and chosen in set by upDiff2.
S66: so circulation, until having chosen set not to be further added by, until there is no any pictures in iterative process
Until element meets above-mentioned condition, gained region is denoted as D2.
S67:D1 is white portion in three components, and part of the region D2 greater than region D1 is grey parts in three components,
Remaining part is divided into the black portions of three components, obtains three components for image segmentation next time.
The foregoing is only a preferred embodiment of the present invention, but scope of protection of the present invention is not limited thereto,
Anyone skilled in the art in the technical scope disclosed by the present invention, according to the technique and scheme of the present invention and its
Inventive concept is subject to equivalent substitution or change, should be covered by the protection scope of the present invention.
Claims (3)
1. a kind of based on the shared sliced image of human body automatic division method for scratching figure and ROI gradual change, it is characterised in that including following
Step:
S1: reading the area-of-interest of image, is scribbled roughly to area-of-interest；
S2: choosing the interior zone A of area-of-interest, calculates interior zone A and the color difference range by scribble pixel, and record is applied
Crow pixel is with the color minus tolerance maximum value loDiff1 of interior zone A pixel and by the color of scribble pixel and interior zone A pixel
Principal-employment maximum value upDiff1；
S3: covering area-of-interest and the borderline region for reading area-of-interest define the sum of interior zone A and borderline region
For region B, region B and the color difference range by scribble pixel are searched, is recorded by the color minus tolerance of scribble pixel and region B pixel most
Big value loDiff2 and by the color principal-employment maximum value upDiff2 of scribble pixel and region B pixel；
S4: by interior zone A be defined as white portion, region B more than region A part be defined as grey parts, B does not include
Region is defined as black portions, obtains whole three component of picture；
S5: the black and white mask of area-of-interest is obtained using shared stingy nomography according to three component information, is calculated with parallel thinning
Method obtains the prospect skeleton of black and white mask；
S6: scribbling prospect skeleton as the prospect of next picture, loDiff1 and upDiff1 according to obtained in S2, S3,
LoDiff2 and upDiff2 data obtain three components of next picture using flooding；
S7: circulation S5 to S6 obtains the black and white mask of all picture area-of-interests until having handled all pictures.
2. a kind of sliced image of human body automatic division method based on shared stingy figure and ROI gradual change according to claim 1,
It is further characterized in that: had according to three component information using the black and white mask that shared stingy nomography obtains area-of-interest in the S5
Body is in the following way:
S51: the maximum quantity parameter K of setting foreground pixel and background pixel extracts K paths from each unknown pixel, if
The angle in each path is π/K, extends unknown pixel region to the rectangular area K × K, wherein the start angle in path is according to the period
Property variation；
S52: it records when encountering foreground pixel or background pixel in each path for the first time, between record foreground pixel and unknown point
Image space and color space distance and the image space between background pixel and unknown point and color space distance, when
Path then stops searching foreground and background pixel beyond the edge of image；
S53: color card is converted by candidate point；Calculate the gap of sample point and unknown point color；Unknown point is calculated to sampling
Foreground point and background dot between straight line path on pixel mutation number；Calculate sampling foreground point and background dot with it is unknown
Physical distance between point；Foreground point and background dot are subjected to the sampled pixel that optimal combination is found in minimum optimum organization；
S54: data processing is carried out to the sampled pixel of optimal combination and obtains the smallest a pair of of prospect background data pair, by the data
To being defined as optimum sampling point；
S55: for the unknown pixel point in three component grey areas, radius threshold is set, to each of in the radius threshold
Unknown pixel point is counted to obtain three minimum MP values, obtains the correlation of the corresponding three unknown pixels point of three minimum MP values
Color data is weighted the relevant colors data and average treatment obtains data T, and data T is defined as optimum sampling
Point；
S56: new foreground pixel, background pixel, transparency and confidence information are obtained according to optimum sampling point and obtain region of interest
The black and white mask in domain.
3. a kind of sliced image of human body automatic division method based on shared stingy figure and ROI gradual change according to claim 1,
It is further characterized in that: three components of next picture are obtained specifically in the following way using flooding in S6:
S61: the location information and color value information of reading prospect skeleton all pixels, setting loDiff1 and upDiff1,
LoDiff2 and upDiff2 value；
S62: the pixel Q in prospect skeleton is chosen, is begun looking for from the neighbor pixel point of four neighborhoods of pixel Q: if working as
The color minus tolerance of preceding pixel point Q and its neighborhood territory pixel is less than loDiff1, and the color of current pixel point Q and its neighborhood territory pixel
Principal-employment is less than upDiff1, then is added to this pixel Q and has chosen in set, until there is no any pixels in iterative process
Until meeting above-mentioned condition, gained region is denoted as region D1；
S63: choosing the pixel W in prospect skeleton, begin looking for from the neighbor pixel point of four neighborhoods of pixel W point, if
The color minus tolerance of current pixel point W and its neighborhood territory pixel is less than loDiff2, and the face of current pixel point W and its neighborhood territory pixel
Color principal-employment is less than upDiff2, then is added to this pixel W and has chosen in set, until there is no any pictures in iterative process
Until element meets above-mentioned condition, gained region is denoted as region D2；
S64: being defined as the part of white portion in three components, region D2 greater than region D1 for region D1 and be defined as grey parts,
Rest part is defined as black portions.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910600636.2A CN110288617B (en) | 2019-07-04 | 2019-07-04 | Automatic human body slice image segmentation method based on shared matting and ROI gradual change |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910600636.2A CN110288617B (en) | 2019-07-04 | 2019-07-04 | Automatic human body slice image segmentation method based on shared matting and ROI gradual change |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110288617A true CN110288617A (en) | 2019-09-27 |
| CN110288617B CN110288617B (en) | 2023-02-03 |
Family
ID=68020587
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910600636.2A Active CN110288617B (en) | 2019-07-04 | 2019-07-04 | Automatic human body slice image segmentation method based on shared matting and ROI gradual change |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110288617B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111462027A (en) * | 2020-03-12 | 2020-07-28 | 中国地质大学（武汉） | Multi-focus image fusion method based on multi-scale gradient and matting |
| CN112101370A (en) * | 2020-11-11 | 2020-12-18 | 广州卓腾科技有限公司 | Automatic pure-color background image matting algorithm, computer-readable storage medium and equipment |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101697229A (en) * | 2009-10-30 | 2010-04-21 | 宁波大学 | Method for extracting region of interest of medical image |
| CN104008547A (en) * | 2014-05-28 | 2014-08-27 | 大连理工大学 | Method for visible serial segmentation of human body slice images based on skeleton angular points |
| CN106875412A (en) * | 2017-02-28 | 2017-06-20 | 重庆理工大学 | A kind of two segmentation localization methods of overlap fruit based on aberration |
| CN106875399A (en) * | 2017-01-04 | 2017-06-20 | 努比亚技术有限公司 | A kind of method for realizing interactive image segmentation, device and terminal |
| CN107452010A (en) * | 2017-07-31 | 2017-12-08 | 中国科学院长春光学精密机械与物理研究所 | A kind of automatically stingy nomography and device |
| CN107730528A (en) * | 2017-10-28 | 2018-02-23 | 天津大学 | A kind of interactive image segmentation and fusion method based on grabcut algorithms |
| CN108564528A (en) * | 2018-04-17 | 2018-09-21 | 福州大学 | A kind of portrait photo automatic background weakening method based on conspicuousness detection |
| CN108986109A (en) * | 2018-06-27 | 2018-12-11 | 大连理工大学 | A kind of serializing viewing human sectioning image automatic division method |
| CN108986107A (en) * | 2018-06-15 | 2018-12-11 | 大连理工大学 | The serializing viewing human sectioning image automatic division method scribbled based on spectrum analysis and skeleton |
-
2019
- 2019-07-04 CN CN201910600636.2A patent/CN110288617B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101697229A (en) * | 2009-10-30 | 2010-04-21 | 宁波大学 | Method for extracting region of interest of medical image |
| CN104008547A (en) * | 2014-05-28 | 2014-08-27 | 大连理工大学 | Method for visible serial segmentation of human body slice images based on skeleton angular points |
| CN106875399A (en) * | 2017-01-04 | 2017-06-20 | 努比亚技术有限公司 | A kind of method for realizing interactive image segmentation, device and terminal |
| CN106875412A (en) * | 2017-02-28 | 2017-06-20 | 重庆理工大学 | A kind of two segmentation localization methods of overlap fruit based on aberration |
| CN107452010A (en) * | 2017-07-31 | 2017-12-08 | 中国科学院长春光学精密机械与物理研究所 | A kind of automatically stingy nomography and device |
| CN107730528A (en) * | 2017-10-28 | 2018-02-23 | 天津大学 | A kind of interactive image segmentation and fusion method based on grabcut algorithms |
| CN108564528A (en) * | 2018-04-17 | 2018-09-21 | 福州大学 | A kind of portrait photo automatic background weakening method based on conspicuousness detection |
| CN108986107A (en) * | 2018-06-15 | 2018-12-11 | 大连理工大学 | The serializing viewing human sectioning image automatic division method scribbled based on spectrum analysis and skeleton |
| CN108986109A (en) * | 2018-06-27 | 2018-12-11 | 大连理工大学 | A kind of serializing viewing human sectioning image automatic division method |
Non-Patent Citations (6)
| Title |
|---|
| HOWARD CHUNG等: "Automated segmentation of muscle and adipose tissue on CT images for human body composition analysis", 《SPIE：MEDICAL IMAGING2009: VISUALIZATION, IMAGE-GUIDED PROCEDURES, AND MODELING》 * |
| 刘茂奇: "基于人体切片的图像分割技术研究", 《中国优秀硕士学位论文全文数据库 (信息科技辑)》 * |
| 孟蕊等: "基于OneCut和共享抠图算法的自适应衣物目标抠取", 《智能计算机与应用》 * |
| 张盛林等: "基于图像抠图技术的多聚焦图像融合方法", 《计算机应用》 * |
| 彭莎等: "基于数字抠图的ROI提取方式改进及其在CT图像中的应用", 《中国生物医学工程学会成立30周年纪念大学暨2010中国生物医学工程学会学术大会论文集》 * |
| 李浪等: "《计算机网络（第二版）》", 31 July 2017 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111462027A (en) * | 2020-03-12 | 2020-07-28 | 中国地质大学（武汉） | Multi-focus image fusion method based on multi-scale gradient and matting |
| CN111462027B (en) * | 2020-03-12 | 2023-04-18 | 中国地质大学（武汉） | Multi-focus image fusion method based on multi-scale gradient and matting |
| CN112101370A (en) * | 2020-11-11 | 2020-12-18 | 广州卓腾科技有限公司 | Automatic pure-color background image matting algorithm, computer-readable storage medium and equipment |
| CN112101370B (en) * | 2020-11-11 | 2021-08-24 | 广州卓腾科技有限公司 | Automatic image matting method for pure-color background image, computer-readable storage medium and equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110288617B (en) | 2023-02-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109242844B (en) | Pancreatic cancer tumor automatic identification system based on deep learning, computer equipment and storage medium | |
| Sun et al. | Saunet: Shape attentive u-net for interpretable medical image segmentation | |
| Zhang et al. | Boundary finding based multi-focus image fusion through multi-scale morphological focus-measure | |
| Wang et al. | Salient object detection based on multi-scale contrast | |
| CN104751187B (en) | Meter reading automatic distinguishing method for image | |
| US9042648B2 (en) | Salient object segmentation | |
| EP3819859B1 (en) | Sky filter method for panoramic images and portable terminal | |
| Li et al. | Continuous depth map reconstruction from light fields | |
| CN111161311A (en) | Visual multi-target tracking method and device based on deep learning | |
| Rhemann et al. | High resolution matting via interactive trimap segmentation | |
| CN103164855B (en) | A kind of Bayesian decision foreground extracting method in conjunction with reflected light photograph | |
| CN109858436A (en) | Target category modification method, detection method based on video dynamic foreground mask | |
| US7840074B2 (en) | Method and apparatus for selecting an object in an image | |
| KR20120061764A (en) | Image processing device, image processing method, and program | |
| CN102034247B (en) | Motion capture method for binocular vision image based on background modeling | |
| EP2869265A1 (en) | Method and apparatus for alpha matting | |
| Jiang et al. | A novel white blood cell segmentation scheme based on feature space clustering | |
| Ofir et al. | On detection of faint edges in noisy images | |
| CN115661144A (en) | Self-adaptive medical image segmentation method based on deformable U-Net | |
| CN105809673B (en) | Video foreground dividing method based on SURF algorithm and the maximum similar area of merging | |
| CN111462027B (en) | Multi-focus image fusion method based on multi-scale gradient and matting | |
| CN110288617A (en) | Based on the shared sliced image of human body automatic division method for scratching figure and ROI gradual change | |
| CN111383759A (en) | Automatic pneumonia diagnosis system | |
| Wang et al. | Adaptive shape prior in graph cut segmentation | |
| Feng et al. | Improved deep fully convolutional network with superpixel-based conditional random fields for building extraction |
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 | ||
| GR01 | Patent grant |