CN104331883A - Image boundary extraction method based on non-symmetry and anti-packing model - Google Patents
Image boundary extraction method based on non-symmetry and anti-packing model Download PDFInfo
- Publication number
- CN104331883A CN104331883A CN201410588458.3A CN201410588458A CN104331883A CN 104331883 A CN104331883 A CN 104331883A CN 201410588458 A CN201410588458 A CN 201410588458A CN 104331883 A CN104331883 A CN 104331883A
- Authority
- CN
- China
- Prior art keywords
- image
- region
- subpattern
- pixel
- boundary
- 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
- G06T9/00—Image coding
- G06T9/40—Tree coding, e.g. quadtree, octree
Abstract
The invention discloses an image boundary extraction method based on a non-symmetry and anti-packing model. The extraction method includes the steps: firstly, performing NAM (non-symmetry and anti-packing model) expression for a binary image to obtain a total sub-pattern number and a coordinate table; secondly, sequentially scanning according to a raster scanning sequence from a first sub pattern at a left upper corner, acquiring relevant parameter values when one sub pattern is scanned, combining areas by a merge-find-set algorithm with union by rank and a path compression strategy if areas of neighboring pixels and an area of the current sub pattern are different and can be combined, otherwise, continuing to scan a next neighboring pixel. Boundary information is updated when neighboring pixels of a left boundary and an upper boundary of the sub pattern are completely scanned and the sub pattern is processed, a next sub pattern is processed according to the steps, and finally the boundary information of the binary image is outputted until all sub patterns are processed. The image boundary extraction method has the advantages of small occupied storage space, high image boundary extraction speed and the like.
Description
Technical field
The present invention relates to a kind of computer image processing technology, particularly one is based on the image boundary extraction method of asymmetric inversed placement model (NAM).
Background technology
Image boundary extraction technology and image representing method closely related, the Boundary Extraction technology of the main hierarchical data structure with regard to image and image introduces relevant present Research and recent development trend both domestic and external here.
(1) hierarchical data structure of image;
Image represents it is one of current research field the most active, and it plays a part very crucial in compression of images, feature extraction, image retrieval, image denoising and the application such as image restoration, image boundary extraction.Effective image represents that algorithm can not only save storage space, but also is conducive to the speed improving image procossing.Existing many bianry images based on spatial data structure represent algorithm at present, as: the expression algorithm of string representation algorithm, tree construction and code word set represent algorithm.With regard to the compression algorithm of bianry image, although the compression performance of compression standard JBIG is always better than any bianry image based on spatial data structure at present and represents algorithm, but because JBIG represents that algorithm relates to entropy code process, for many application being the JBIG form that impossible operate compression.Hierarchical data structure is very important region representation method in the fields such as computer vision, robot, computer graphics, image procossing, pattern-recognition.Quadtrees (QT, Quad Tree) is a kind of form of image layered expression, it be study the earliest, be also a kind of layer representation form studying at most.Early stage quadtrees represents it is all quadtrees structure based on pointer, and in order to significantly reduce storage space, the people such as Gargantini eliminate pointer scheme, propose the method for expressing being referred to as linear quadtree (LQT, Linear Quad Tree).The people such as Subramanian have studied the image representing method based on space binary tree segmentation (BSP, Binary Space Partitioning).Image is after BSP tree representation, and it represents that result directly can support the compression of image and segmentation scheduling algorithm.Represent based on the binary tree mixed and quadtrees, the people such as Kassim propose a kind of image representing method based on layering segmentation.Based on the coding method of B-tree bougainvillea shape, the gray level image that the people such as Distasi propose based on spatial data structure represents algorithm.Based on S-data tree structure and Gouraud shadowing method, the gray level image that the people such as Chung propose a kind of spatial data structure based on S-tree represents (STC, S-Tree Coding) method (K.L.Chung, J.G.Wu.Improved image compression using S-tree and shading approach.IEEE Transactions on Communications, 2000,48 (5): 748-751.).Subsequently, the people such as Chung propose a kind of mixing gray scale image table based on DCT domain and spatial domain and show (SDCT, Spatial-and DCT-based) method (K.L.Chung, Y.W.Liu, W.M.Yan.A hybrid gray image representation using spatial-and DCT-based approach with application to moment computation.Journal of Visual Communication and Image Representation, 2006,17 (6): 1209-1226).
Although above-mentioned individual-layer data indicates many advantages, they too emphasize the symmetry split, and are not therefore optimum method for expressing.By means of the thought of Packing problem, to find the maximized asymmetric dividing method of segmentation for target, the people such as Chen propose the NAM method for expressing of image model.Zheng Yun equality people proposes a kind of based on the Gouraud shadowing method of expansion and the NAM gray level image method for expressing of non-overlapped rectangle subpattern, referred to as RNAMC method for expressing (Zheng Yunping, Chen Chuanbo. a kind of new gray level image represents algorithm research. Chinese journal of computers, 2010,33 (12): 2397-2406.).Due to can overlapping NAM method for expressing generally can be higher than non-overlapped NAM method for expressing efficiency, Zheng Yun equality people also been proposed a kind of based on the Gouraud shadowing method of expansion and the NAM gray level image method for expressing of overlapping rectangles subpattern, referred to as ORNAM method for expressing (Yunping Zheng, Zhiwen Yu, Jane You, Mudar Sarem.A novel gray image representation using overlapping rectangular NAM and extended shading approach.Journal of Visual Communication and Image Representation, 2012, 23 (7): 972-983.).Experimental result shows: compare with RNAMC method for expressing with STC, SDCT, under the prerequisite keeping picture quality, ORNAM method for expressing has higher ratio of compression and less block number, thus more effectively can reduce data space, is a kind of good method that gray level image represents.Recently, Zheng Yun equality also been proposed a kind of NAM method for expressing of new bianry image and applies it in areal calculation, achieve good result (Yunping Zheng, Mudar Sarem.A novel binary image representation algorithm by using NAM and coordinate encoding procedure and its application to area calculation.Frontiers of Computer Science, 2014,8 (5): 763-772.).From present situation, LQT represents the complicacy that mainly concentrates on and reduce image-processing operations and to the expansion of wider scope, theoretic achievement is a lot, applies to actual quite a few, and more and more, be still a kind of popular image representing method in current image processing field.
Image representing method has two objects: the first, improves the expression efficiency of image.The second, improve the processing speed of image manipulation.
(2) the Boundary Extraction technology of image;
The profile of border normally object, can be people and describes or identify that target and interpretation of images provide vital feature.Therefore, the border in image identified and extracts, playing an important role in computer vision and digital image analysis and application, also there is important practical value.For many years, image boundary detection and extraction are the important research themes in Digital Image Processing, analysis and validation field always.Traditional boundary detection method has derivative method, gradient method, Laplce's method and variously to improve one's methods.In recent years, multi-scale morphology, based on the rim detection of mathematical morphology with have also been obtained application by fuzzy logic to the technology that image boundary detects.In essence, traditional boundary detection method is the method based on pixel grey scale change.Be generally the state first detecting each pixel and its neighbor, to determine whether this pixel is on the border of object, then represent Boundary Detection image with the gray-scale value of pixel in image or with two-value gray level image.Traditional boundary detect to be boundary pixel point with the key of extracting method detect performance and frontier point join algorithm performance.In the Boundary Detection application of complicated image, effect is often undesirable.In recent years, the application of object-oriented image analysis method receives publicity gradually.With traditional difference based on pixel grey scale disposal route be, by Iamge Segmentation being several mutually not overlapping regions (image object), subsequently using image object as fundamental analysis processing unit; This mode, relative to using pixel as basic processing unit, be more suitable in conjunction with the cognitive knowledge of the mankind about real world, thus more effectively can extract the image-region be consistent with real world target (atural object) in shape and classification from image.This object-based graphical analysis is a kind of new theory occurred in recent years.By using the BSP method of hierarchy, the people such as Wang, C.C.L. propose a kind of efficient BSP solid boundaries extracting method based on shearing manipulation.Their polygon algorithm repeats the trimming operation on the body cell of the corresponding convex division in space, connects computation bound by traversal cell.They use point-based representations to be similar to (Wang, C.C.L. along with BSP solid boundaries is raised the efficiency and generated to finite-precision arithmetic; Manocha, D., Efficient Boundary Extraction of BSP Solids Based on Clipping Operations.IEEE Transactions on Visualization and Computer Graphics, 2013,19 (1): 16-29).
In sum, although there is a lot of researchist to be devoted to the research of image boundary extraction in recent years, propose much new Boundary Extraction technology, due to the difficulty of problem itself, mostly current method is for certain specific tasks, goes back the solution that neither one is general.The important reason of of image boundary extraction difficulty is complicacy and the diversity of image.Due to the complicacy of image, the independent image boundary extraction algorithm of existing any one is all difficult to obtain gratifying segmentation result to general pattern, thus while continuing to be devoted to that new concept, new theory, new method are introduced image boundary extraction field, more pay attention to effective combination of multiple boundary extraction algorithm, the method great majority proposed in recent years combine many algorithms, compared with single boundary extraction algorithm, Boundary Extraction integrated technology is more effective, and robustness, stability, accuracy and adaptivity etc. are better.
Effective image representing method can not only save storage space, and can also improve the speed of image procossing.The NAM method for expressing of image model is against layout expression way to the one of image model, image model is expressed as predefined subpattern set in essence, subpattern can be stored, therefore the method also directly supports the Processing Algorithm such as the Boundary Extraction of image.
Invention Inner holds
The object of the invention is for Problems existing in conventional images Boundary Extraction technology, there is provided a kind of image boundary extraction method based on asymmetric inversed placement model (NAM), this image boundary extraction method can significantly improve expression and the extraction efficiency of image boundary extraction.First boundary extraction algorithm based on NAM will encode to image, obtains total subpattern number n, coordinates table W after encoding.Then from the subpattern of first, the upper left corner, press the order scanning subpattern of raster scanning successively, often scan the coordinate figure that a subpattern just obtains these subpattern four corners from coordinates table, then the west circle of this subpattern and all neighbor pixels of northern boundary are scanned, if the region belonging to neighbor pixel and current subpattern do not belong to same region and can merge, then perform the Union-find Sets algorithm combined region that band presses order merging and path compression strategy, otherwise continue the next neighbor pixel of scanning.When the left margin of this subpattern and the neighbor pixel of coboundary all scanned, this subpattern is disposed, upgrade boundary information, by the next subpattern of above step process, until all subpattern process complete the boundary information that can extract bianry image.
Object of the present invention is achieved through the following technical solutions: a kind of image boundary extraction method based on asymmetric inversed placement model, comprises the following steps:
Step S1, use the image b that size is G × H by the bianry image representation based on asymmetric inversed placement model to encode, obtain total subpattern number n, coordinates table W after encoding.
Concrete method for expressing is as follows:
Step S1.1, be that the size of 0, M is equal with pending bianry image b by all elements assignment of matrix variables M, be G × H, with the counting variable n=0 of seasonal subpattern; Wherein, G and H is natural number;
Starting point (the x of step S1.2, determine in bianry image b by the order of raster scanning one not identified rectangle subpattern
1, y
1), determine according to this starting point the subpattern that an area is maximum, and subpattern is made a check mark in bianry image b;
The parameter of step S1.3, record subpattern, that is: the coordinate (x in the upper left corner
1, y
1), the coordinate (x in the lower right corner
2, y
2); Make n=n+1;
Step S1.4, circulation perform step (S1.2) to (S1.3), until the subpattern in bianry image b is all identified complete;
Step S1.5, according to following coordinate data compression algorithm, the coordinate of nonzero elements all in matrix variables M to be encoded, and coding result is stored in a coordinates table W;
1. size of lining by line scan is the matrix variables M of G × H, if this row all elements is zero, so just need not to encode this row, in this case, use bit " 0 " to represent that one's own profession does not exist nonzero element from the beginning to the end, and this binary digit " 0 " is stored in the coding schedule W of this row; Otherwise, if there is nonzero element in this row, before each nonzero element, so just add prefix symbol " 1 ", then after prefix symbol, add the code word identifying nonzero element 1,2 and-1, finally this prefix symbol " 1 " and code word are thereafter stored in the coding schedule W of this row;
2. represent the position of this nonzero element column with x bit, and by this x bit storage in the coding schedule W of this row, wherein the value of x calculates by following two kinds of situations;
For first nonzero element run in certain a line, x=[log
2h]; Here x bit is used for indicating the position of first nonzero element about one's own profession head end;
For other nonzero elements except first nonzero element run in certain a line, x=[log
2(H-c)], wherein c is the position of the row of the front nonzero element once run into; Here x bit is used for representing the position of this nonzero element about the right-hand member of the nonzero element of front first encoding;
3. after last nonzero element of certain a line has been encoded, bit " 0 " is used to represent that the remaining element of one's own profession is zero, and this binary digit " 0 " is stored in the coding schedule W of this row, otherwise, if the position of last nonzero element of this row is at the end of one's own profession, so just need not use " 0 " to represent the remaining element of one's own profession to be zero;
Step S1.6, output coordinate table W, wherein W is linked in sequence by the row coding schedule of all row of matrix variables M and obtains.
Step S2, put the sequence number j of a Current Scan subpattern, and make j=0, arrange a pointer matrix B, size is G × H simultaneously, for representing the region that each pixel is pointed to.
Step S3, in coordinates table, obtain W [j].
Step S4, according to W [j], calculate size size and left margin, the coboundary coordinate information of current subpattern.
Step S5, from left margin bottom, up scan, each left margin pixel L is found out to a pixel LL (namely LL is less than L by 1 in X-direction) on its left side, and utilize matrix B to find out region belonging to pixel L and pixel LL, to merge by order with band again and the Union-find Sets algorithm of path compression strategy finds out the ancestors region in these two regions, if two regions are the same areas, then jump to next pixel, otherwise, if two regions do not belong to the same area, then judge whether these two ancestors can merge according to average and variance.
2 data structures used in this step and next step are as follows:
1. Region data structure field:
{Mean，Var，Size，Father，Count，EdgeLink}
Wherein Mean represents the gray average in this region, and Var represents the gray variance in this region, and Size represents the size in this region, i.e. pixel count, and these three territories are used for the union operation of support area.Territory Father is a pointer, is used to refer to the father node to this region, and Count is used for the quantity of descendent areas in this region, and above two territories are used for supporting Union-find Sets algorithm.Territory EdgeLink points to the border in this region, can be used for following the trail of the boundary information in this region.
2. Edge data structure field:
{PreLink，First，Last，SucLink}
PreLink and SucLink is used for supporting doubly linked list, First and Last points to starting point and the terminal on summit, corner.
After step S6, left margin are scanned, from the leftmost of coboundary, to turn right scanning, one, its top pixel TT (namely TT is less than T by 1 in the Y direction) is found out to each coboundary pixel T, and utilize matrix B to find out region belonging to pixel T and pixel TT, to merge by order with band again and the Union-find Sets algorithm of path compression strategy finds out the ancestors region in these two regions, if two regions are the same areas, then jump to next pixel; Otherwise, if two regions do not belong to the same area, then judge whether these two ancestors can merge according to average and variance.
Step S7, renewal boundary information, j++, jumps to step S3, until all subpatterns are disposed.
The boundary information of step S8, output bianry image b.
Principle of the present invention: the present invention is by means of the thought of the quadtrees region representation method of location problem and bianry image, basis is expressed as by the NAM of bianry image, provide a kind of image boundary extraction method based on asymmetric inversed placement model, the method can significantly improve expression and the operating efficiency of image boundary extraction.First image boundary extraction method based on asymmetric inversed placement model will carry out NAM expression to bianry image, obtains total subpattern number n and coordinates table W.Then from the subpattern of first, the upper left corner, press the order scanning of raster scanning successively, often scan a subpattern and just obtain related parameter values, if the region belonging to neighbor pixel and current subpattern do not belong to same region and can merge, then perform the Union-find Sets algorithm combined region that band presses order merging and path compression strategy, otherwise continue the next neighbor pixel of scanning.When the left margin of this block and the neighbor pixel of coboundary all scanned, this block is disposed, and upgrades boundary information, by the next subpattern of above step process, until all subpattern process complete, finally exports the boundary information of bianry image.The time complexity of the inventive method is O (nL α (n)), and wherein n represents the block number of homogeneous blocks, and L represents the length of side size of every block, and α (n) is the inverse function of ackerman function.
The present invention compared to existing technology tool has the following advantages:
1, in the expression of image boundary extraction, for 4 given width images, the subpattern number average out to 11585 (36633) of NAM (LQT), the compression ratio CR average out to 1.3801 (0.2862) of NAM (LQT), also namely the ratio of compression of RNAM is 4.8222 times of the ratio of compression of LQT.NAM represents than LQT, algorithm also represents that algorithm decreases 68.38% in the number of subpattern simultaneously, obviously, NAM is better than LQT in the number (decreasing 68.38%) and ratio of compression increase rate (improve 382.22%) of subpattern.
2, in the speed of image boundary extraction, the execution speed of the Boundary Extraction represented based on NAM on average improves 92.39% than the execution speed of the Boundary Extraction represented based on LQT, because of but a kind of more effective boundary extraction algorithm, have that to take storage space little, image boundary extraction speed is fast.
Therefore, NAM boundary extraction method provided by the invention is better than LQT boundary extraction method.The present invention both can be applicable to traditional image boundary extraction market, can be applicable to emerging field again, as network transmission, wireless telecommunications, medical image etc.
3, the present invention is compared with the existing boundary extraction method based on LQT, boundary extraction method based on NAM has lower bit rate and less subpattern number, thus more effectively can reduce the speed of data space and raising image boundary extraction, because of but the better boundary extraction method of the one of bianry image, this boundary extraction method can be applied to the various aspects of image procossing, at reduction storage space, accelerates to have good theoretical reference meaning and actual application value in transmission speed, raising pattern match efficiency etc.
Accompanying drawing explanation
Fig. 1 is the entire flow figure of the boundary extraction method that the present invention is based on NAM.
Fig. 2 is the method for expressing process flow diagram of the bianry image based on NAM.
Fig. 3 a is the standard bianry image F16 of the present invention's 512 × 512 sizes used.
Fig. 3 b is the standard bianry image Goldhill of the present invention's 512 × 512 sizes used.
Fig. 3 c is the standard bianry image Lena of the present invention's 512 × 512 sizes used.
Fig. 3 d is the standard bianry image Peppers of the present invention's 512 × 512 sizes used.
Fig. 4 a gives the segmentation effect of the LQT method of Fig. 3 c.
Fig. 4 b gives the effect of the region merging technique of LQT method.
Fig. 4 c gives the border of the bianry image extracted by LQT method.
Fig. 5 a gives the segmentation effect of the LQT method of Fig. 3 c.
Fig. 5 b gives the effect of the region merging technique of NAM method.
Fig. 5 c gives the border of the bianry image extracted by NAM method.
Embodiment
Embodiment
The object of the invention is for Problems existing in conventional images Boundary Extraction technology, a kind of image boundary extraction method based on asymmetric inversed placement model (NAM) is provided, its overall procedure as shown in Figure 1, can significantly improve the Boundary Extraction speed of image and effectively can reduce storage space simultaneously.First image boundary extraction method based on asymmetric inversed placement model will carry out NAM expression to bianry image, obtains total subpattern number n and coordinates table W.Then from the subpattern of first, the upper left corner, press the order scanning of raster scanning successively, often scan a subpattern and just obtain related parameter values, if the region belonging to neighbor pixel and current subpattern do not belong to same region and can merge, then perform the Union-find Sets algorithm combined region that band presses order merging and path compression strategy, otherwise continue the next neighbor pixel of scanning.When the left margin of this block and the neighbor pixel of coboundary all scanned, this block is disposed, and upgrades boundary information, by the next subpattern of above step process, until all subpattern process complete, finally exports the boundary information of bianry image.The time complexity of the inventive method is O (nL α (n)), and wherein n represents the block number of homogeneous blocks, and L represents the length of side size of every block, and α (n) is the inverse function of ackerman function.Experimental result shows: compared with the current popular boundary extraction method based on LQT, the boundary extraction method based on NAM that the present invention proposes has lower bit rate and less subpattern number, thus more effectively can reduce data space and improve the speed of image boundary extraction, because of but the better boundary extraction method of the one of bianry image.This method can be applied to the various aspects of image procossing, at reduction storage space, accelerates to have good theoretical reference meaning and actual application value in transmission speed, raising pattern match efficiency etc.
As shown in Figure 2, image representing method provided by the invention is represented with rectangle NAM by the bianry image b being G × H to a given width size, obtain set and a coordinates table W of mutually different subpattern, then based on these subpatterns, a kind of image boundary extraction method based on asymmetric inversed placement model is proposed.Specifically comprise the following steps:
(S1) the image b that to use based on the bianry image representation of asymmetric inversed placement model be G × H by size encodes, and obtains total subpattern number n, coordinates table W after encoding.
Concrete method for expressing is as follows:
(S1.1) be that the size of 0, M is equal with pending bianry image b by all elements assignment of matrix variables M, be G × H, with the counting variable n=0 of seasonal subpattern; Wherein, G and H is natural number;
(S1.2) starting point (x of one that determines in bianry image b by the order of raster scanning not identified rectangle subpattern
1, y
1), determine according to this starting point the subpattern that an area is maximum, and subpattern is made a check mark in bianry image b;
(S1.3) parameter of subpattern is recorded, that is: the coordinate (x in the upper left corner
1, y
1), the coordinate (x in the lower right corner
2, y
2); Make n=n+1;
(S1.4) circulation performs step (S1.2) to (S1.3), until the subpattern in bianry image b is all identified complete;
(S1.5) according to following coordinate data compression algorithm, the coordinate of nonzero elements all in matrix variables M is encoded, and coding result is stored in a coordinates table W;
1. size of lining by line scan is the matrix variables M of G × H, if this row all elements is zero, so just need not to encode this row, in this case, use bit " 0 " to represent that one's own profession does not exist nonzero element from the beginning to the end, and this binary digit " 0 " is stored in the coding schedule W of this row; Otherwise, if there is nonzero element in this row, before each nonzero element, so just add prefix symbol " 1 ", then after prefix symbol, add the code word identifying nonzero element 1,2 and-1, finally this prefix symbol " 1 " and code word are thereafter stored in the coding schedule W of this row;
2. represent the position of this nonzero element column with x bit, and by this x bit storage in the coding schedule W of this row, wherein the value of x calculates by following two kinds of situations;
For first nonzero element run in certain a line, x=[log
2h]; Here x bit is used for indicating the position of first nonzero element about one's own profession head end;
For other nonzero elements except first nonzero element run in certain a line, x=[log
2(H-c)], wherein c is the position of the row of the front nonzero element once run into; Here x bit is used for representing the position of this nonzero element about the right-hand member of the nonzero element of front first encoding;
3. after last nonzero element of certain a line has been encoded, bit " 0 " is used to represent that the remaining element of one's own profession is zero, and this binary digit " 0 " is stored in the coding schedule W of this row, otherwise, if the position of last nonzero element of this row is at the end of one's own profession, so just need not use " 0 " to represent the remaining element of one's own profession to be zero;
(S1.6) output coordinate table W, wherein W is linked in sequence by the row coding schedule of all row of matrix variables M and obtains.
(S2) put the sequence number j of a Current Scan subpattern, and make j=0, arrange a pointer matrix B simultaneously, size is G × H, for representing the region that each pixel is pointed to.
(S3) in coordinates table, obtain W [j].
(S4) according to W [j], size size and left margin, the coboundary coordinate information of current subpattern is calculated.
(S5) from left margin bottom, up scan, each left margin pixel L is found out to a pixel LL (namely LL is less than L by 1 in X-direction) on its left side, and utilize matrix B to find out region belonging to pixel L and pixel LL, to merge by order with band again and the Union-find Sets algorithm of path compression strategy finds out the ancestors region in these two regions, if two regions are the same areas, then jump to next pixel, otherwise, if two regions do not belong to the same area, then judge whether these two ancestors can merge according to average and variance.
2 data structures used in this step and next step are as follows:
1. Region data structure field:
{Mean，Var，Size，Father，Count，EdgeLink}
Wherein Mean represents the gray average in this region, and Var represents the gray variance in this region, and Size represents the size in this region, i.e. pixel count, and these three territories are used for the union operation of support area.Territory Father is a pointer, is used to refer to the father node to this region, and Count is used for the quantity of descendent areas in this region, and above two territories are used for supporting Union-find Sets algorithm.Territory EdgeLink points to the border in this region, can be used for following the trail of the boundary information in this region.
2. Edge data structure field:
{PreLink，First，Last，SucLink}
PreLink and SucLink is used for supporting doubly linked list, First and Last points to starting point and the terminal on summit, corner.
(S6) left margin scanned after, from the leftmost of coboundary, to turn right scanning, one, its top pixel TT (namely TT is less than T by 1 in the Y direction) is found out to each coboundary pixel T, and utilize matrix B to find out region belonging to pixel T and pixel TT, to merge by order with band again and the Union-find Sets algorithm of path compression strategy finds out the ancestors region in these two regions, if two regions are the same areas, then jump to next pixel; Otherwise, if two regions do not belong to the same area, then judge whether these two ancestors can merge according to average and variance.
(S7) upgrade boundary information, j++, jumps to (S3), until all subpatterns are disposed.
(S8) boundary information of bianry image b is exported.
To LQT and NAM these two kinds, this example represents that algorithm compares.The size of the test pattern adopted, title are respectively the 4 width gray level images such as ' F16 ', ' Goldhill ', ' Lena ' and ' Peppers ' of 512 × 512 sizes, as shown in Fig. 3 a, Fig. 3 b, Fig. 3 c and Fig. 3 d.
This example first gives the comparison of the experimental result of NAM and LQT method for expressing, and the expression efficiency of these 2 kinds of methods can be measured by following 2 parameters, that is: the number of subpattern and ratio of compression.Table 1 (table 1 is the comparison sheet of compression performance) gives NAM and represents that algorithm and LQT represent the comparison of algorithm on compression performance.Table 2 (table 2 is the comparison sheet of subpattern number and block number) gives NAM and LQT and represents the comparison of algorithm on the number of subpattern number and block number.
| Algorithm | F16 | Goldhill | Lena | Pepper |
| LQT | 0.3244 | 0.2060 | 0.2720 | 0.3425 |
| NAM | 1.7443 | 0.8786 | 1.2486 | 1.6488 |
Table 1
| Algorithm | F16 | Goldhill | Lena | Pepper |
| LQT | 31078 | 48946 | 37072 | 29437 |
| NAM | 8396 | 17203 | 11374 | 9365 |
Table 2
As can be seen from table 1 and 2, for 4 given width images, the CR average out to 1.3801 (0.2862) of NAM (LQT), the subpattern average out to 11585 (36633) of NAM (LQT), also namely the ratio of compression of NAM is 4.8222 times of the ratio of compression of LQT, thus is more conducive to saving storage space; And NAM represents also represent decreased average 68.38% than LQT in the number of subpattern, thus be more conducive to improving the speed of efficiency that image represents and image procossing.
In sum, compared with representing algorithm with LQT, NAM represents that algorithm has higher ratio of compression and less block number, thus more effectively can reduce data space, because of but the better method for expressing of the one of bianry image.
As shown in Fig. 4 a, Fig. 4 b and Fig. 4 c, be the segmentation of the LQT of ' Lena ', merging and Boundary Extraction result; As shown in Fig. 5 a, Fig. 5 b and Fig. 5 c, be the segmentation of the NAM of ' Lena ', merging and Boundary Extraction result.Fig. 4 is an example of LQT method, and wherein, Fig. 4 a gives the segmentation effect of the LQT method of Fig. 3 c, and Fig. 4 b gives the effect of the region merging technique of LQT method, and Fig. 4 c gives the border of the bianry image extracted by LQT method.Fig. 5 is an example of NAM method, and wherein, Fig. 5 a gives the segmentation effect of the LQT method of Fig. 3 c, and Fig. 5 b gives the effect of the region merging technique of NAM method, and Fig. 5 c gives the border of the bianry image extracted by NAM method.
Shown in following table 3 (table 3 represent the segmentation performance comparison sheet represented with LQT for NAM), for NAM represents comparing of the segmentation efficiency represented with LQT.
| Image | F16 | Goldhill | Lena | Peppers |
| NP | 262144 | 262144 | 262144 | 262144 |
| NB LQT | 31078 | 48946 | 37072 | 29437 |
| NB NAM | 8396 | 17203 | 11374 | 9365 |
| Nreg LQT | 575 | 2405 | 1788 | 1628 |
| Nreg NAM | 575 | 2405 | 1788 | 1628 |
| T LQT | 1406 | 2688 | 1875 | 1343 |
| T NAM | 94 | 156 | 141 | 140 |
| (T LQT-T NAM)/T LQT | 93.31％ | 94.20％ | 92.48％ | 89.58％ |
Table 3
In table 3, NP represents the sum of all pixels of test pattern, and unit is individual; NB
lQTand NB
nAMrepresent that image represents the quantity of subpattern when representing with NAM with LQT respectively, unit is individual; Nreg
lQTand Nreg
nAMthe quantity in the region after splitting when representing that image represents represent with NAM with LQT respectively, unit is individual; T
nAMand T
lQTrepresent the execution time of the partitioning algorithm represented based on LQT and represent based on NAM respectively, unit is millisecond.
For 4 given width images, the CR average out to 1.3801 (0.2862) of NAM (LQT), the subpattern average out to 11585 (36633) of NAM (LQT), also namely the ratio of compression of NAM is 4.8222 times of the ratio of compression of LQT, thus is more conducive to saving storage space; And NAM represents also represent decreased average 68.38% than LQT in the number of subpattern, thus be more conducive to improving the speed of efficiency that image represents and image procossing.
In addition, be not difficult to find out from table 3, the execution speed of the partitioning algorithm represented based on NAM on average improves 92.39% than the execution speed of the partitioning algorithm represented based on LQT, because of but a kind of more effective partitioning algorithm.
In sum, the result of the present embodiment confirms the correctness of analysis.
Above-described embodiment is the present invention's preferably embodiment; but embodiments of the present invention are not restricted to the described embodiments; change, the modification done under other any does not deviate from Spirit Essence of the present invention and principle, substitute, combine, simplify; all should be the substitute mode of equivalence, be included within protection scope of the present invention.
Claims (3)
1., based on an image boundary extraction method for asymmetric inversed placement model, it is characterized in that, comprise the following steps:
Step S1, use the image b that size is G × H by the bianry image representation based on asymmetric inversed placement model to encode, obtain total subpattern number n, coordinates table W after encoding;
Step S2, put the sequence number j of a Current Scan subpattern, and make j=0, arrange a pointer matrix B, size is G × H simultaneously, for representing the region that each pixel is pointed to;
Step S3, in coordinates table, obtain W [j];
Step S4, according to W [j], calculate size size and left margin, the coboundary coordinate information of current subpattern;
Step S5, from left margin bottom, up scan, each left margin pixel L is found out to a pixel LL on its left side, that is: LL is less than L by 1 in X-direction, and utilize matrix B to find out region belonging to pixel L and pixel LL, to merge by order with band again and the Union-find Sets algorithm of path compression strategy finds out the ancestors region in these two regions, if two regions are the same areas, then jump to next pixel, otherwise, if two regions do not belong to the same area, then judge whether these two ancestors can merge according to average and variance;
After step S6, left margin are scanned, from the leftmost of coboundary, to turn right scanning, one, its top pixel TT is found out to each coboundary pixel T, that is: TT is less than T by 1 in the Y direction, and utilizes matrix B to find out region belonging to pixel T and pixel TT, then to merge by order with band and the Union-find Sets algorithm of path compression strategy finds out the ancestors region in these two regions, if two regions are the same areas, then jump to next pixel; Otherwise, if two regions do not belong to the same area, then judge whether these two ancestors can merge according to average and variance;
Step S7, renewal boundary information, j++, jumps to step S3, until all subpatterns are disposed;
The boundary information of step S8, output bianry image b.
2. the image boundary extraction method based on asymmetric inversed placement model according to claim 1, it is characterized in that, described step 1 comprises the following steps:
Step S1.1, be that the size of 0, M is equal with pending bianry image b by all elements assignment of matrix variables M, be G × H, with the counting variable n=0 of seasonal subpattern; Wherein, G and H is natural number;
Starting point (the x of step S1.2, determine in bianry image b by the order of raster scanning one not identified rectangle subpattern
1, y
1), determine according to this starting point the subpattern that an area is maximum, and subpattern is made a check mark in bianry image b;
The parameter of step S1.3, record subpattern, that is: the coordinate (x in the upper left corner
1, y
1), the coordinate (x in the lower right corner
2, y
2); Make n=n+1;
Step S1.4, circulation perform step (S1.2) to (S1.3), until the subpattern in bianry image b is all identified complete;
Step S1.5, according to coordinate data compression algorithm, the coordinate of nonzero elements all in matrix variables M to be encoded, and coding result is stored in a coordinates table W;
Step S1.6, output coordinate table W, wherein W is linked in sequence by the row coding schedule of all row of matrix variables M and obtains.
3. the image boundary extraction method based on asymmetric inversed placement model according to claim 1, is characterized in that, employs Region data structure field and Edge data structure field in step S5 and step S6;
Described Region data structure field is: Mean, Var, Size, Father, Count, EdgeLink},
Wherein, Mean represents the gray average in this region, and Var represents the gray variance in this region, and Size represents the size in this region, i.e. pixel count, and these three territories are used for the union operation of support area; Territory Father is a pointer, is used to refer to the father node to this region, and Count is used for the quantity of descendent areas in this region, and above two territories are used for supporting Union-find Sets algorithm; Territory EdgeLink points to the border in this region, is used for following the trail of the boundary information in this region;
Described Edge data structure field is: PreLink, First, Last, SucLink},
Wherein, PreLink and SucLink is used for supporting doubly linked list, First and Last points to starting point and the terminal on summit, corner.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410588458.3A CN104331883B (en) | 2014-10-28 | 2014-10-28 | A kind of image boundary extraction method based on asymmetric inversed placement model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410588458.3A CN104331883B (en) | 2014-10-28 | 2014-10-28 | A kind of image boundary extraction method based on asymmetric inversed placement model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104331883A true CN104331883A (en) | 2015-02-04 |
| CN104331883B CN104331883B (en) | 2017-06-06 |
Family
ID=52406603
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410588458.3A Expired - Fee Related CN104331883B (en) | 2014-10-28 | 2014-10-28 | A kind of image boundary extraction method based on asymmetric inversed placement model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104331883B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109993762A (en) * | 2019-02-21 | 2019-07-09 | 华南理工大学 | A kind of rapid edge-detection method based on asymmetric inverse layout pattern |
| CN110798332A (en) * | 2018-08-03 | 2020-02-14 | Emc Ip控股有限公司 | Method and system for searching directory access group |
| CN111060116A (en) * | 2019-12-04 | 2020-04-24 | 江西洪都航空工业集团有限责任公司 | Grassland self-drawing system based on vision |
| CN113706639A (en) * | 2021-07-21 | 2021-11-26 | 国网江苏省电力有限公司电力科学研究院 | Image compression method and device based on rectangular NAM, storage medium and computing equipment |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102572431A (en) * | 2011-12-29 | 2012-07-11 | 华南理工大学 | Overlapping rectangular subpattern-based non-symmetry and anti-packing model (NAM) image representation method |
| CN102629387A (en) * | 2012-02-29 | 2012-08-08 | 华南理工大学 | Image geometry moment processing method based on non-symmetry and anti-packing pattern representation model |
-
2014
- 2014-10-28 CN CN201410588458.3A patent/CN104331883B/en not_active Expired - Fee Related
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102572431A (en) * | 2011-12-29 | 2012-07-11 | 华南理工大学 | Overlapping rectangular subpattern-based non-symmetry and anti-packing model (NAM) image representation method |
| CN102629387A (en) * | 2012-02-29 | 2012-08-08 | 华南理工大学 | Image geometry moment processing method based on non-symmetry and anti-packing pattern representation model |
Non-Patent Citations (5)
| Title |
|---|
| WU XUELI等: "《A New Fast Edge Detection Algorithm》", 《GEO-SPATIAL INFORMATION SCIENCE》 * |
| 吴雪丽: "《基于斜面分解的非对称逆布局图像表示方法与处理算法研究》", 《中国博士学位论文全文数据库 信息科技辑》 * |
| 吴雪丽等: "《一种快速边缘检测算法》", 《小型微型计算机系统》 * |
| 彭俊杰: "《基于图的NAM表示及其上的显著性区域检测》", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 费峥峰等: "《一种快速彩色图像颜色分割算法》", 《微计算机信息》 * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110798332A (en) * | 2018-08-03 | 2020-02-14 | Emc Ip控股有限公司 | Method and system for searching directory access group |
| CN110798332B (en) * | 2018-08-03 | 2022-09-06 | Emc Ip控股有限公司 | Method and system for searching directory access groups |
| US11500873B2 (en) | 2018-08-03 | 2022-11-15 | EMC IP Holding Company LLC | Methods and systems for searching directory access groups |
| CN109993762A (en) * | 2019-02-21 | 2019-07-09 | 华南理工大学 | A kind of rapid edge-detection method based on asymmetric inverse layout pattern |
| CN109993762B (en) * | 2019-02-21 | 2022-12-16 | 华南理工大学 | Rapid edge detection method based on asymmetric inverse layout mode |
| CN111060116A (en) * | 2019-12-04 | 2020-04-24 | 江西洪都航空工业集团有限责任公司 | Grassland self-drawing system based on vision |
| CN111060116B (en) * | 2019-12-04 | 2023-07-18 | 江西洪都航空工业集团有限责任公司 | Independent grassland map building system based on vision |
| CN113706639A (en) * | 2021-07-21 | 2021-11-26 | 国网江苏省电力有限公司电力科学研究院 | Image compression method and device based on rectangular NAM, storage medium and computing equipment |
| CN113706639B (en) * | 2021-07-21 | 2024-02-23 | 国网江苏省电力有限公司电力科学研究院 | Image compression method and device based on rectangular NAM, storage medium and computing equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104331883B (en) | 2017-06-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102930561B (en) | Delaunay-triangulation-based grid map vectorizing method | |
| CN102496023B (en) | Region of interest extraction method of pixel level | |
| CN103337072B (en) | A kind of room objects analytic method based on texture and geometric attribute conjunctive model | |
| CN104318560B (en) | A kind of image partition method based on asymmetric inversed placement model | |
| Schick et al. | An evaluation of the compactness of superpixels | |
| CN104331883A (en) | Image boundary extraction method based on non-symmetry and anti-packing model | |
| CN102800094A (en) | Fast color image segmentation method | |
| CN108805863B (en) | Method for detecting image change by combining deep convolutional neural network with morphology | |
| CN101840582B (en) | Boundary digitizing method of cadastral plot | |
| CN106874818B (en) | A kind of Digital Object Unique Identifier DOI recognition methods and device | |
| CN105931180A (en) | Salient information guided image irregular mosaic splicing method | |
| CN105068918B (en) | A kind of page method of testing and device | |
| CN104637066B (en) | The quick framework extraction method of bianry image based on sequential refinement | |
| CN103049340A (en) | Image super-resolution reconstruction method of visual vocabularies and based on texture context constraint | |
| CN110399760A (en) | A kind of batch two dimensional code localization method, device, electronic equipment and storage medium | |
| CN103679808A (en) | Method and system for rebuilding three-dimensional head model by two-dimensional nuclear magnetic images | |
| CN105516540A (en) | Compression method and device of binary image | |
| CN102542528B (en) | Image conversion processing method and system | |
| CN105354570A (en) | Method and system for precisely locating left and right boundaries of license plate | |
| CN101364306B (en) | Stationary image compression coding method based on asymmetric inversed placement model | |
| CN107832732A (en) | Method for detecting lane lines based on ternary tree traversal | |
| US9875555B2 (en) | Partitioning an image | |
| CN102270338B (en) | Method for effectively segmenting repeated object based on image representation improvement | |
| Allende et al. | A non-parametric filter for digital image restoration, using cluster analysis | |
| CN102799884B (en) | Image target outline extracting method based on Gestalt rule |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20170606 Termination date: 20211028 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |