CN105654457A - Device and method for processing image - Google Patents
Device and method for processing image Download PDFInfo
- Publication number
- CN105654457A CN105654457A CN201410648544.9A CN201410648544A CN105654457A CN 105654457 A CN105654457 A CN 105654457A CN 201410648544 A CN201410648544 A CN 201410648544A CN 105654457 A CN105654457 A CN 105654457A
- Authority
- CN
- China
- Prior art keywords
- image
- mark
- representative
- image region
- representativeness
- 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
Landscapes
- Image Analysis (AREA)
Abstract
The invention relates to a device and a method for processing an image. The device comprises an acquisition unit used for acquiring the image including at least one object, a segmentation unit used for carrying out segmentation for the image to acquire multiple image subareas, a calculation unit used for calculating representative marks of the image, a mapping unit used for mapping each of the representative marks into one of the multiple image subareas, and a counting unit used for counting the image subareas having mapping relationship with the representative marks to acquire the quantity of the objects in the image, wherein at least two of the representative marks are mapped to one same image subarea of the multiple image subareas, a local extremum area of the image is calculated by the counting unit on the basis of a pixel value of each pixel point, and the local extremum areas are taken as the representative marks. Through the device, the quantity of the objects in the image can be calculated, and the external contours of the objects can be segmented.
Description
Technical field
It relates to the technical field of image procossing, specifically relate to the device for being counted by the object in image and split and method.
Background technology
This part provides the background information relevant with the disclosure, and this is prior art not necessarily.
Usually cultivate the objects such as some bacteriums, red corpuscle, white corpuscle or crystallization in the medium, and often need to be counted by these objects and split in order to further biological analysis. In recent years, in order to the objects such as the bacterium in image, red corpuscle, white corpuscle or crystallization are counted, adopt artificial counting method more, first the image of the objects such as bacterium, red corpuscle, white corpuscle or crystallization is obtained by microscope, then carry out artificial counting, have the grid of fixed size in tally, it may be achieved stdn, but this method operation is loaded down with trivial details, time-consuming, and counting precision is low, error is big. Simultaneously, object number in these substratum is various, and there is the phenomenon that some objects are sticked together, and the Object Segmentation adopting traditional dividing method these to be sticked together is an object, cause segmentation inaccurate, thus affect follow-up research and analytical work.
Summary of the invention
This part provides general summary of the present disclosure, instead of comprehensive disclosure of its whole scope or its whole feature.
Object of the present disclosure is to provide a kind of image processing apparatus and image processing method, it can by counting, to the image region that the representative mark with objects in images has mapping relation, the number obtaining objects in images, and by optionally non-representative area being merged into the exterior contour figure that representative area obtains object, it is thus possible to comparatively accurately calculate the number of objects in images efficiently, and it is partitioned into the exterior contour of object.
According to one side of the present disclosure, it provides a kind of image processing apparatus, this device comprises: acquiring unit, and it comprises the image of at least one object for obtaining; Cutting unit, it is for carrying out over-segmentation to described image, to obtain multiple image region; Calculating unit, it is for calculating the representative mark of described image; Map unit, it is for mapping each in described representativeness mark in described multiple image region, and at least two representative marks in wherein said representativeness mark map to the same image region in described multiple image region; And counting unit, it is for counting marking the image region with mapping relation with described representativeness, with the number of described object obtained in described image, wherein, described calculating unit is based on the pixel value of each pixel in described image, calculate the local extremal region of described image, using described local extremal region as described representativeness mark.
According to another aspect of the present disclosure, it provides a kind of image processing method, the method comprises: obtain the image comprising at least one object;Described image is carried out over-segmentation, to obtain multiple image region; Calculate the representative mark of described image; Mapping each in described representativeness mark in described multiple image region, at least two representative marks in wherein said representativeness mark map to the same image region in described multiple image region; And count marking the image region with mapping relation with described representativeness, with the number of described object obtained in described image, wherein, the representative mark calculating described image comprises: based on the pixel value of each pixel in described image, calculate the local extremal region of described image, using described local extremal region as described representativeness mark.
According to another aspect of the present disclosure, provide a kind of machinable medium, it carries the program product comprising the machine readable instructions code being stored therein, wherein, described instruction code is when being read by computer and perform, it is possible to described computer is performed according to image processing method of the present disclosure.
Use according to image processing apparatus of the present disclosure and method, image is carried out over-segmentation, obtain multiple image region, and the representative mark of computed image, each in representativeness being marked maps in multiple image region, and wherein at least two representative marks map to same image region. Thus, can count, to marking with representativeness the image region with mapping relation, the number obtaining objects in images such that it is able to more accurately calculate the number of objects in images efficiently than traditional method according to image processing apparatus of the present disclosure and method. Further, use according to image processing apparatus of the present disclosure and method, the image region with mapping relation will be marked as the representative area of object with representativeness, and optionally non-representative area except representative area in multiple image region is merged into representative area to obtain the exterior contour figure of object such that it is able to accurately it is partitioned into the exterior contour of object. Simultaneously, adopt image processing apparatus and the method for the present invention, when the object in image not being split, object can be counted roughly, and accuracy is higher than traditional method of counting, thus when only needing to be counted by the object in image, it is possible to omit the step split by object, it is achieved simple, time saving and energy saving, computation complexity is low.
From, the description provided at this, further suitability region will become obvious. Description and specific examples in this summary are the object in order to illustrate, and are not intended to restriction the scope of the present disclosure.
Accompanying drawing explanation
The accompanying drawing described here enforcement that just not all is possible in order to the object of signal of selected embodiment, and it is not intended to restriction the scope of the present disclosure. In the accompanying drawings:
Fig. 1 be technical scheme of the present disclosure for the schematic diagram of the example object in image being counted and splitting;
Fig. 2 is the block diagram of the image processing apparatus according to embodiment of the present disclosure;
Image is carried out over-segmentation to obtain the schematic diagram of multiple image region according to embodiment of the present disclosure by Fig. 3;
Fig. 4 is the schematic diagram of the representative mark of the computed image according to embodiment of the present disclosure;
Fig. 5 maps to the schematic diagram of in multiple image region according to each in representativeness being marked of embodiment of the present disclosure;
Fig. 6 is the block diagram of the image processing apparatus according to another embodiment of the present disclosure;
Fig. 7 is the block diagram of the image processing apparatus according to another embodiment of the present disclosure;
Fig. 8 is the block diagram of the image processing apparatus according to another embodiment of the present disclosure;
Fig. 9 is the schematic diagram of the minimum enclosed rectangle of the representative area according to embodiment of the present disclosure;
Non-representative area is optionally merged into representative area to obtain the schematic diagram of the exterior contour figure of object according to embodiment of the present disclosure by Figure 10;
Figure 11 is the schema of the image processing method according to embodiment of the present disclosure;
Image in Fig. 1 is carried out the effect schematic diagram after smoothing processing according to embodiment of the present disclosure by Figure 12;
Image in Figure 12 is carried out over-segmentation to obtain the effect schematic diagram of multiple image region according to embodiment of the present disclosure by Figure 13;
Figure 14 is the representative effect schematic diagram marked of the image calculating in Figure 12 according to embodiment of the present disclosure;
The representative mark calculated in Figure 14 is mapped the effect schematic diagram to the image region in Figure 10 according to embodiment of the present disclosure by Figure 15;
Figure 16 is the effect schematic diagram of the exterior contour figure of the object of the image obtaining in Fig. 1 according to embodiment of the present disclosure; And
Figure 17 is the block diagram of the example arrangement of the general purpose personal computer that wherein can realize the image processing apparatus according to embodiment of the present disclosure and method.
Although the disclosure easily stands various amendment and replacement form, but its specific embodiment is shown in the drawings as an example, and describes in detail at this. However it is understood that at this description of specific embodiment do not intend to be restricted to the disclosure disclosed specific form, but on the contrary, disclosure object to be covered all modifications, the equivalence that drop within spirit and scope of the present disclosure and replace. It is noted that run through several accompanying drawings, corresponding label indicates corresponding parts.
Embodiment
With reference now to accompanying drawing, example of the present disclosure is described more fully. Hereinafter describe substantially just exemplary, and it is not intended to the restriction disclosure, application or purposes.
Provide example embodiment, so that the disclosure will become detailed, and will fully pass on its scope to those skilled in the art. Set forth numerous specific detail such as the example of particular elements, device and method, to provide the detailed understanding to embodiment of the present disclosure. Will being apparent that to those skilled in the art, it is not necessary to use specific details, example embodiment can be implemented by many different forms, and they should not be interpreted as restriction the scope of the present disclosure. In some example embodiment, it does not have describe well-known process, well-known structure and well-known technology in detail.
Fig. 1 show technical scheme of the present disclosure for the schematic diagram of the example object in image split and counts. As shown in Figure 1, in the image obtained by microscope, there are some objects in irregular shape, wherein may there is inter-adhesive phenomenon in multiple object, and that is, multiple object may contact with each other or part is handed over folded. Meanwhile, the object number in image is various, not of uniform size, and shape is irregular, thus counting and segmentation are caused certain difficulty. Object in image is such as but not limited to bacterium, red corpuscle, white corpuscle or crystallization etc.
As mentioned above, adopt traditional counting and dividing method, operate loaded down with trivial details, time-consuming, and counting is low with segmentation precision, error big, thus affect follow-up research and analytical work. Use technical scheme of the present disclosure, the pixel value of the pixel of object region intermediate can be utilized higher, the feature that the pixel value of the pixel of borderline region is lower carrys out the representative mark of computed image, and is counted by the image region being marked with mapping relation with representativeness thus obtain the number of object. Further, it is also possible to obtained the exterior contour figure of object by the mode of merging image region, thus realize the object of object count and segmentation. , it may also be useful to technical scheme of the present disclosure, it is not necessary to first the object in image being carried out segmentation can count, it is achieved simple, in addition the complexity of counting is reduced.
Fig. 2 shows the block diagram of the image processing apparatus according to embodiment of the present disclosure.Image processing apparatus 200 according to embodiment of the present disclosure can comprise acquiring unit 210, cutting unit 220, calculating unit 230, map unit 240 and counting unit 250.
Acquiring unit 210 comprises the image of at least one object for obtaining. In one embodiment, the image comprising at least one object can be the image obtained by microscope. Object in image includes but not limited to bacterium, red corpuscle, white corpuscle or crystallization etc., and it is inter-adhesive to there are at least two objects in image, namely contacts with each other or partly hands over folded. In one embodiment, the image of acquisition is sent to cutting unit 220 and calculates unit 230 by acquiring unit 210 respectively.
Cutting unit 220 for image is carried out over-segmentation, to obtain multiple image region. In one embodiment, cutting unit 220 can obtain from acquiring unit 210 needs the image carrying out over-segmentation. Herein, it is possible to adopt dividing method well known in the art that image is carried out over-segmentation, such as, based on the dividing method of figure, mean-shift algorithm and quick-shift algorithm etc. Over-segmentation can regard a kind of method of coarse segmentation as, object is when retaining the boundary information of objects in images as far as possible, it is multiple less image regions by Iamge Segmentation, the phenomenon of over-segmentation may occur in this process, the region segmentation being about to originally belong to same target has become multiple image region, is therefore called over-segmentation. In one embodiment, the image after over-segmentation can be sent to map unit 240 by cutting unit 220.
Calculate the representative mark of unit 230 for computed image. Representative mark is the local extremal region of image. In one embodiment, calculate unit 230 and obtain the image needing to calculate representative mark from acquiring unit 210, and the pixel value based on each pixel in image, the local extremal region calculating described image is as described representativeness mark. In the picture, generally, the pixel value of the pixel of object region intermediate is higher, and the pixel value of the pixel of borderline region is lower, and this feature thus can be utilized to carry out the representative mark of computed image. Such as, the pixel value of each pixel in image being calculated, when the pixel value of all pixels in certain region is all higher than the pixel value of the pixel in its adjacent domain, this region is representative mark. In one image, it is possible to there is one or more representative mark.
In one embodiment, owing to an object in image in theory has a representative mark. Therefore, when needing the object in image carries out rough counting, it is possible to the representative mark calculated by calculating unit is counted, to obtain the number of objects in images. But, in the image that reality obtains, due to the complicacy of objects in images, may there is multiple representativeness mark in an object, thus in one embodiment, if needing to be counted comparatively accurately by the object in image, then calculate unit 230 and the representative mark of the image calculated can be sent to map unit 240.
Map unit 240 is for mapping each in representativeness mark in multiple image region. In one embodiment, map unit 240 can obtain multiple image regions of image from cutting unit 220, and obtain the representative mark of image from calculating unit 230.Each representative mark is mapped to an image region by map unit 240, that is, for each representative mark, a unique image region is had to have mapping relation with it, and for each image region, can there is one or more representative mark and there is with it mapping relation, it is also possible to there is not representative mark and there is with it mapping relation, the mapping relation of " multipair " namely between representative mark and image region, may be there is. In one embodiment, at least two representative marks in representative mark map to the same image region in multiple image region. In one embodiment, representativeness can be marked the mapping relation between image region and/or mark the image region with mapping relation with representativeness and be sent to counting unit 250 by map unit 240.
Counting unit 250 for the image region having a mapping relation with representativeness mark is counted, with the number of object obtained in image. In one embodiment, counting unit 250 can obtain the image region with representative mark with mapping relation from map unit 240, and it is counted. In another embodiment, counting unit 250 can obtain the mapping relation between representative mark and image region from map unit 240, thus obtains the image region with representative mark with mapping relation, and it is counted. In one embodiment, counting unit 250 can the number of object in output image.
Utilize the object region intermediate pixel value in image higher as mentioned above, it is necessary, calculate unit 230, the representative mark of the feature calculating object that borderline region pixel value is lower. Therefore, when image also exists the object of some adhesions, it is also possible to distinguish these objects by representativeness mark, thus realize the counting to object. Meanwhile, representativeness mark is mapped to image region by map unit 240, and counting unit 250 counts marking the image region with mapping relation with representativeness, instead of is counted by representativeness mark. Therefore, when there is multiple local extremal region in certain object in image, namely during multiple representativeness mark, this representativeness mark can not be repeated counting by counting unit such that it is able to more accurately determine the number of objects in images.
Fig. 3 shows and image carries out over-segmentation to obtain the schematic diagram of multiple image region according to embodiment of the present disclosure. As shown in Figure 3, image is carried out over-segmentation and obtain image region 1-9. For convenience of explanation, Fig. 3 illustrate only 9 image regions, and the boundary line between region is straight line, in the image over-segmentation of reality, it is possible to exist more than 9 or less than the image region of 9, and the boundary line between region can be curve.
Fig. 4 shows the schematic diagram of the representative mark of the computed image according to embodiment of the present disclosure. As shown in Figure 4, image exists representative mark A-G. For convenience of explanation, Fig. 4 illustrate only 7 representative marks, in the image of reality, it is possible to exist more than 7 or less than the representative mark of 7, and the shape of representative mark can be more irregular.
Fig. 5 shows and maps to the schematic diagram of in multiple image region according to each in representativeness being marked of embodiment of the present disclosure. In embodiments, representativeness is marked the representative mark mapping of each in A-G to an image region in image region 1-9 according to the rule reserved in advance by map unit 240.
In one embodiment, when representativeness mark is covered by multiple image region, this representativeness mark is mapped in the plurality of image region by map unit 240.
Fig. 6 shows the block diagram of the image processing apparatus according to another embodiment of the present disclosure. As shown in Figure 6, image processing apparatus 200 also comprises the first judging unit 270, and this first judging unit 270 can judge that representativeness marks whether by a covering in multiple image region, and judged result is sent to map unit 240. Map unit 240 obtains judged result from the first judging unit 270, and performs to operate accordingly. Herein, cover and represent that a region has hidden another region completely. Representative mark is covered by an image region, and namely all parts in the region of representative mark are all coated in the region of this image region. Such as, as shown in Figure 5, representative mark A and B is covered by image region 1, and representative mark C is covered by image region 2, and representative mark G is covered by image region 4. In the present embodiment, representativeness is marked A and B and all maps image region 1 by map unit 240, representativeness marks C and maps image region 2, and representativeness marks G and maps image region 4.
In one embodiment, when representativeness mark overlaps mutually with the more than one image region in multiple image region, this representativeness mark is mapped to an image region in this more than one image region by map unit 240. In one embodiment, the first judging unit 270 can judge that representativeness marks whether to overlap mutually with the more than one image region in multiple image region, and judged result is sent to map unit 240. Herein, hand over a folded expression region to have together with part area coincidence with another region, between two regions, there is not the relation that covers another. Such as, as shown in Figure 5, representative mark D and image region 3, image region 4 and image region 9 overlap mutually, and representative mark E overlaps mutually with image region 1 and image region 2, and representative flag F overlaps mutually with image region 8 and image region 9. In the present embodiment, representativeness is marked D and maps in image region 3, image region 4 and image region 9 by map unit 240, representativeness is marked E and maps in image region 1 and image region 2, and representative flag F is mapped in image region 8 and image region 9.
In the above-described embodiment, when there is the image region not covering other representative mark in the middle of more than one image region, map unit 240 chooses the maximum image region of the area overlapped mutually with this representativeness mark in the middle of the image region not covering other representative mark, as marking the image region with mapping relation with this representativeness; When in the middle of more than one image region, each image region is covered with other representative mark, map unit 240 chooses the maximum image region of the area overlapped mutually with this representativeness mark in the middle of more than one image region, as marking the image region with mapping relation with this representativeness. In one embodiment, the first judging unit 270 can judge whether there is the image region not covering other representative mark in the middle of more than one image region, and judged result is sent to map unit 240.Such as, as shown in Figure 5, D is marked for representativeness, image region 4 has covered other representative mark G, image region 3 and image region 9 do not cover other representative mark, and therefore map unit 240 chooses the maximum image region 3 of the area overlapped mutually with representativeness mark D as the image region with it with mapping relation in these two image regions; E is marked for representativeness, image region 1 covers other representative mark A and B, image region 2 covers other representative mark C, and therefore map unit 240 chooses the maximum image region 2 of the area overlapped mutually with representativeness mark E as the image region with it with mapping relation in these two image regions; For representative flag F, image region 8 and image region 9 all do not cover other representative mark, and therefore map unit 240 chooses the maximum image region 8 of the area overlapped mutually with representative flag F as the image region with it with mapping relation in these two image regions.
According to above-mentioned enforcement mode, in the example as shown in fig. 5, representativeness is marked A-G and maps image region 1,1,2,3,2,8 and 4 respectively by map unit 240. It thus is seen that representative mark A and B is mapped to same image region 1, representative mark C and E is mapped to same image region 2. In the present embodiment, representativeness mark is mapped to image region by map unit 240 according to the rule of above-mentioned setting, and in actual applications, representativeness mark can also be mapped to image region by map unit 240 according to other rules. Then, mapping result can be sent to counting unit 250 by map unit 240. In the example as shown in fig. 5, counting unit 250 counts marking the image region 1,2,3,4 and 8 with mapping relation with representativeness, and the number obtaining objects in images is 5.
Fig. 7 shows the block diagram of the image processing apparatus according to another embodiment of the present disclosure. As shown in Figure 7, image processing apparatus 200 also comprises merge cells 260.
Merge cells 260 for marking the image region with mapping relation as the representative area of object with representativeness. In one embodiment, merge cells 260 can obtain the image region with representative mark with mapping relation from map unit 240, and is defined as representative area. In another embodiment, merge cells 260 can obtain the mapping relation of representative mark and image region from map unit 240, thus obtains the image region with representative mark with mapping relation, and is defined as representative area. In one embodiment, image region except representative area in multiple image region can also be defined as non-representative area by merge cells 260. Such as, in the example as shown in fig. 5, image region 1,2,3,4 and 8 is defined as the representative area of object by merge cells 260, and image region 5,6,7 and 9 is defined as non-representative area.
Merge cells 260 also for optionally non-representative area being merged into representative area, to obtain the exterior contour figure of object. In the example as shown in fig. 5, non-representative area 5,6,7 and 9 is merged into representative area 1,2,3,4 and 8 according to pre-set rules selection by merge cells 260, to obtain the exterior contour figure of object. In one embodiment, merge cells 260 can the exterior contour figure of object in output image.
As mentioned above, it is necessary, non-representative area is optionally merged into representative area thus obtains the exterior contour figure of objects in images by merge cells 260 such that it is able to more accurately determine the profile of objects in images. Simultaneously, adopt image processing apparatus and the method for the present invention, when the object in image not being split, object can be counted roughly, and accuracy is higher than traditional method of counting, thus when only needing to be counted by the object in image, it is possible to omit the step split by object, it is achieved simple, time saving and energy saving, computation complexity is low.
In one embodiment, when non-representative area overlaps mutually with in representative mark, the representative area with mapping relation that this non-representative area is merged in marking with representativeness by merge cells 260; When non-representative area overlaps mutually with multiple in representative mark, merge cells 260 chooses a maximum representative mark of the area overlapped mutually with non-representative area in multiple representativeness mark, and this non-representative area is merged into the representative mark with this and has the representative area of mapping relation.
Fig. 8 shows the block diagram of the image processing apparatus according to another embodiment of the present disclosure. As shown in Figure 8, image processing apparatus 200 also comprises the 2nd judging unit 280,2nd judging unit 280 can judge whether representative area overlaps mutually with representativeness mark, and representative area is in marking with representativeness or multiple overlaps mutually, and judged result is sent to merge cells 260. Merge cells 260 obtains judged result from the 2nd judging unit 280, and performs to operate accordingly. Such as, as shown in Figure 5, non-representative area 9 overlaps mutually with representative mark D and representative flag F, thus merge cells 260 marks in representativeness in D and representative flag F and chooses the maximum representative flag F of the area overlapped mutually with non-representative area 9, and is merged into by non-representative area 9 and has in the representative area 8 of mapping relation with representative flag F.
In one embodiment, when non-representative area is covered by the minimum enclosed rectangle of in representative area, or when the ratio of the area of the area that the minimum enclosed rectangle of in representative area and this non-representative area overlap mutually and this non-representative area is greater than predetermined threshold, merge cells 260 one that this non-representative area is merged in this representative area. In one embodiment, 2nd judging unit 280 can judge whether non-representative area is covered by the minimum enclosed rectangle of in representative area, and whether there is a representative area, make the ratio of the area of area that the minimum enclosed rectangle of in this representative area and non-representative area overlap mutually and this non-representative area be greater than predetermined threshold, and judged result is sent to merge cells 260. In the present embodiment, all representative areas all can be made its minimum enclosed rectangle, then the 2nd judging unit 280 determines whether this minimum enclosed rectangle covers one or more non-representative area, or the ratio of the area of the area whether overlapped mutually with certain non-representative area and this non-representative area is greater than predetermined threshold. Herein, predetermined threshold can be set according to the demand of algorithm and processing accuracy by technician, and such as 90%, 95% etc.In the present embodiment, it is possible to adopting algorithm well known in the art to calculate the minimum enclosed rectangle of representative area, this is not limited by the present invention.
Fig. 9 shows the schematic diagram of the minimum enclosed rectangle of the representative area according to embodiment of the present disclosure. As shown in Figure 9, the minimum enclosed rectangle of representative area 1 covers non-representative area 5, and therefore in the present embodiment, non-representative area 5 is merged in representative area 1 by merge cells 260.
According to above-mentioned enforcement mode, in the example as shown in fig. 5, non-representative area 9 is merged in representative area 8 by merge cells 260, and is merged in representative area 1 by non-representative area 5, to obtain the exterior contour figure of objects in images. In the above-described embodiment, non-representative area is merged into representative area according to above-mentioned default rules selection by merge cells 260, and in actual applications, non-representative area can also be merged into representative area according to other rules selections by merge cells 260.
Figure 10 shows and non-representative area is optionally merged into representative area to obtain the schematic diagram of the exterior contour figure of object according to embodiment of the present disclosure, and as shown in Figure 10, merge cells 260 obtains the exterior contour figure of 5 objects in image.
In one embodiment, image processing apparatus 200 also comprises processing unit 290, before image carrying out over-segmentation at cutting unit 220 and calculates the representative mark of unit 230 computed image, image is carried out smoothing processing. In one embodiment, processing unit 290 can obtain the image comprising at least one object from acquiring unit 210, and the image after smoothing processing is sent to cutting unit 220 and calculates unit 230. Herein, it is possible to use image is carried out smoothing processing by Processing Algorithm as known in the art, such as, being made the Morphological scale-space method etc. of original image smoothing by " opening reconstruction " and " closing reconstruction ", this is not limited by the present invention. After smoothing processing, it is possible to the impact of noise in removal of images, and in image, create smooth region.
Below in conjunction with Figure 11, the image processing method according to embodiment of the present disclosure is described. As shown in figure 11, image processing method according to embodiment of the present disclosure starts from step S1110. In step S1110, obtain the image comprising at least one object.
Next, in step S1120, image is carried out over-segmentation, to obtain multiple image region.
Next, in step S1130, the representative mark of computed image. Wherein, the representative mark of computed image comprises: based on the pixel value of each pixel in image, the local extremal region of computed image, by this local extremal region representatively property mark.
Next, in step S1140, each in representativeness being marked maps in multiple image region, and wherein at least two representative marks in representative mark map to the same image region in multiple image region.
Next, in step S1150, the image region having a mapping relation with representativeness mark is counted, with the number of described object obtained in image.
According to embodiment of the present disclosure, this image processing method comprises further and will mark the image region with mapping relation with representativeness as the representative area of object; Optionally non-representative area except representative area in multiple image region is merged into representative area, to obtain the exterior contour figure of object.
According to embodiment of the present disclosure, optionally non-representative area except representative area in multiple image region is merged into representative area to comprise: when non-representative area overlaps mutually with in representative mark, this non-representative area is merged into and representativeness mark in a representative area with mapping relation.
According to embodiment of the present disclosure, optionally non-representative area except representative area in multiple image region is merged into representative area to comprise: when non-representative area overlaps mutually with multiple in representative mark, multiple representativeness mark is chosen the representative mark that the area overlapped mutually with non-representative area is maximum, and this non-representative area is merged into the representative mark with this there is the representative area of mapping relation.
According to embodiment of the present disclosure, optionally non-representative area except representative area in multiple image region is merged into representative area to comprise: when non-representative area is covered by the minimum enclosed rectangle of in representative area, or when the ratio of the area of the area that the minimum enclosed rectangle of in representative area and non-representative area overlap mutually and this non-representative area is greater than predetermined threshold, by one that this non-representative area is merged in this representative area.
According to embodiment of the present disclosure, each mapping in representativeness being marked comprises in multiple image region: when representativeness mark is covered by multiple image region, map this representativeness mark in the plurality of image region.
According to embodiment of the present disclosure, each mapping in representativeness being marked comprises in multiple image region: when representativeness mark overlaps mutually with the more than one image region in multiple image region, map this representativeness mark to an image region in this more than one image region.
According to embodiment of the present disclosure, this representativeness mark is mapped and comprises to an image region in this more than one image region: when there is the image region not covering other representative mark in the middle of more than one image region, the image region that the area overlapped mutually with this representativeness mark is maximum is chosen, as marking the image region with mapping relation with this representativeness in the middle of the image region not covering other representative mark; And when in the middle of more than one image region, each image region is covered with other representative mark, the image region that the area overlapped mutually with this representativeness mark is maximum is chosen, as marking the image region with mapping relation with this representativeness in the middle of more than one image region.
According to embodiment of the present disclosure, comprised further before the representative mark that image is carried out over-segmentation and computed image and image is carried out smoothing processing.
Did description in detail before the various embodiments of the above-mentioned steps of the image processing method according to embodiment of the present disclosure, it be not repeated.
Use according to image processing apparatus of the present disclosure and method, first image is carried out over-segmentation, obtain multiple image region, and the representative mark of computed image, then each in representativeness being marked maps in multiple image region, and wherein at least two representative marks map to same image region. Thus, can count, to marking with representativeness the image region with mapping relation, the number obtaining objects in images such that it is able to more accurately calculate the number of objects in images efficiently than traditional method according to image processing apparatus of the present disclosure and method. Further, use according to image processing apparatus of the present disclosure and method, the image region with mapping relation will be marked as the representative area of object with representativeness, and optionally non-representative area is merged into representative area to obtain the exterior contour figure of object such that it is able to accurately it is partitioned into the exterior contour of object. Simultaneously, adopt image processing apparatus and the method for the present invention, when the object in image not being split, object can be counted roughly, and accuracy is higher than traditional method of counting, thus when only needing to be counted by the object in image, it is possible to omit the step split by object, it is achieved simple, time saving and energy saving, computation complexity is low.
The effect of the present invention is described below in conjunction with the effect schematic diagram obtained according to embodiment of the present disclosure.Figure 12 shows, according to embodiment of the present disclosure, image in Fig. 1 carries out the effect schematic diagram after smoothing processing. As shown in figure 12, after carrying out smoothing processing, eliminate the noise in image so that image is more smooth. Figure 13 shows, according to embodiment of the present disclosure, image in Figure 12 is carried out over-segmentation to obtain the effect schematic diagram of multiple image region. As shown in figure 13, after carrying out over-segmentation, image is divided in order to multiple image region. Figure 14 shows the representative effect schematic diagram marked of the image calculating in Figure 12 according to embodiment of the present disclosure. As shown in figure 14, the image of Figure 12 exists four representative marks. Figure 15 shows, according to embodiment of the present disclosure, the representative mark calculated in Figure 14 is mapped the effect schematic diagram to the image region in Figure 13. As shown in figure 15, representative for four in Figure 14 mark is mapped to three image regions in Figure 13, thus the object in image can be counted. Figure 16 shows the effect schematic diagram of the exterior contour figure of the object of the image obtaining in Fig. 1 according to embodiment of the present disclosure. As shown in figure 16, after non-representative area is optionally merged into representative area, obtain the exterior contour figure of three objects in image.
As shown in figures 12-16, image processing apparatus and the method for the present invention is adopted, it is possible to more accurately calculate the number of objects in images than traditional method efficiently, and obtain the exterior contour figure of object, it is achieved simple, error is less.
Obviously, can realize in the way of the computer executable program being stored in the storage media that various machine can read according to each operating process of image processing method of the present disclosure.
And, object of the present disclosure can also be realized by following mode: the storage media storing above-mentioned executable program code is supplied to system or equipment directly or indirectly, and computer in this system or equipment or central processing unit (CPU) read and perform said procedure code. Now, as long as this system or equipment have the function of steering routine, then embodiment of the present disclosure is not limited to program, and this program can also be arbitrary form, such as, target routine, explain the program that device performs or be supplied to the shell script etc. of operating system.
These machinable mediums above-mentioned include but not limited to: various storer and storage unit, semiconductor devices, and disk unit is light, magnetic and magneto-optic disk such as, and other is suitable for the medium etc. of storage information.
In addition, the corresponding website of computer by being connected on Internet, and will download according to computer program code of the present disclosure and be installed in computer then to perform this program, it is also possible to realize technical scheme of the present disclosure.
Figure 17 is the block diagram of the example arrangement of the general purpose personal computer that wherein can realize the image processing apparatus according to embodiment of the present disclosure and method.
As shown in figure 17, CPU1701 performs various process according to the program stored in read-only storage (ROM) 1702 or from storing the program that part 1708 is loaded into random access memory (RAM) 1703. In RAM1703, also store the data required when CPU1701 performs various process etc. as required. CPU1701, ROM1702 and RAM1703 are connected to each other via bus 1704. Input/output interface 1705 is also connected to bus 1704.
Following parts are connected to input/output interface 1705: importation 1706 (comprising keyboard, mouse etc.), output part 1707 (comprise indicating meter, such as cathode tube (CRT), liquid-crystal display (LCD) etc., and loud speaker etc.), store part 1708 (comprising hard disk etc.), communications portion 1709 (comprising NIC such as LAN card, modulator-demodulator unit etc.).Communications portion 1709 performs communication process via network such as Internet. As required, driving mechanism 1710 also can be connected to input/output interface 1705. Detachable media 1711 such as disk, CD, magneto-optic disk, semiconductor memory etc. are installed on driving mechanism 1710 as required so that the computer program therefrom read is installed to as required and stores in part 1708.
When series of processes above-mentioned by software simulating, from network such as Internet or storage media, such as detachable media 1711 installs the program forming software.
It will be understood by one skilled in the art that this kind of storage media be not limited to shown in Figure 17 wherein have program stored therein and equipment distributes the detachable media 1711 to provide program to user separately. The example of detachable media 1711 comprises disk (comprising floppy disk (registered trademark)), CD (comprising cd-rom (CD-ROM) and digital universal disc (DVD)), magneto-optic disk (comprising mini-disk (MD) (registered trademark)) and semiconductor memory. Or, storage media can be ROM1702, store in part 1708 hard disk that comprises etc., wherein computer program stored, and is distributed to user together with comprising their equipment.
In system and method for the present disclosure, it is clear that each parts or each step can decompose and/or again combine. These decompose and/or combination should be considered as equivalents of the present disclosure again. Further, the step performing above-mentioned series of processes can naturally according to the order illustrated temporally order execution, but not need necessarily to perform according to time sequence. Some step can perform parallel or independently of one another.
Although describing embodiment of the present disclosure above by reference to the accompanying drawings in detail, it should be appreciated that, enforcement mode described above is just for illustration of the disclosure, and does not form restriction of the present disclosure. For a person skilled in the art, it is possible to above-mentioned enforcement mode is made various changes and modifications and do not deviate from essence of the present disclosure and scope. Therefore, the scope of the present disclosure is only limited by appended claim and equivalents thereof.
About the enforcement mode comprising above embodiment, following attached note is also disclosed:
Attached note 1. 1 kinds of image processing apparatus, comprising:
Acquiring unit, it comprises the image of at least one object for obtaining;
Cutting unit, it is for carrying out over-segmentation to described image, to obtain multiple image region;
Calculating unit, it is for calculating the representative mark of described image;
Map unit, it is for mapping each in described representativeness mark in described multiple image region, and at least two representative marks in wherein said representativeness mark map to the same image region in described multiple image region; And
Counting unit, it is for counting the image region having a mapping relation with described representativeness mark, with the number of described object obtained in described image,
Wherein, described calculating unit, based on the pixel value of each pixel in described image, calculates the local extremal region of described image, using described local extremal region as described representativeness mark.
The device of attached note 2. according to attached note 1, comprises further:
Merge cells, it for marking the image region with mapping relation as the representative area of described object with described representativeness, optionally non-representative area except described representative area in described multiple image region is merged into described representative area, to obtain the exterior contour figure of described object.
The device of attached note 3. according to attached note 2, wherein, when in described non-representative area and described representativeness mark one is overlapped mutually, described non-representative area is merged into and the described representative area with mapping relation in described representativeness mark by described merge cells.
The device of attached note 4. according to attached note 2, wherein, described non-representative area and described representativeness mark in multiple overlap mutually time, described merge cells chooses a maximum representative mark of the area overlapped mutually with described non-representative area in described multiple representativeness mark, and described non-representative area is merged into the representative area with a described representative mark with mapping relation.
The device of attached note 5. according to attached note 2 or 3, wherein, when described non-representative area is covered by the minimum enclosed rectangle of in described representative area, or when the ratio of the area of the area that the minimum enclosed rectangle of in described representative area and described non-representative area overlap mutually and described non-representative area is greater than predetermined threshold, described merge cells described that described non-representative area is merged in described representative area.
The device of attached note 6. according to attached note 1, wherein, when described representativeness mark is covered by described multiple image region, described representativeness mark is mapped to described in described multiple image region by described map unit.
The device of attached note 7. according to attached note 1 or 6, wherein, when described representativeness mark overlaps mutually with the more than one image region in described multiple image region, described representativeness mark is mapped to an image region in described more than one image region by described map unit.
The device of attached note 8. according to attached note 7, wherein,
When there is the image region not covering other representative mark in the middle of described more than one image region, described map unit chooses the maximum image region of the area overlapped mutually with described representativeness mark in the middle of described image region not covering other representativeness mark, marks the image region with mapping relation as with described representativeness; And
When in the middle of described more than one image region, each image region is covered with other representative mark, described map unit chooses the maximum image region of the area overlapped mutually with described representativeness mark in the middle of described more than one image region, marks the image region with mapping relation as with described representativeness.
The device of attached note 9. according to attached note 1, comprises processing unit further, and it for carrying out smoothing processing to described image before the representative mark that described image is carried out over-segmentation by described cutting unit and described calculating unit calculates described image.
The device of attached note 10. according to attached note 1, wherein, described to liking bacterium.
Attached note 11. 1 kinds of image processing methods, comprising:
Obtain the image comprising at least one object;
Described image is carried out over-segmentation, to obtain multiple image region;
Calculate the representative mark of described image;
Mapping each in described representativeness mark in described multiple image region, at least two representative marks in wherein said representativeness mark map to the same image region in described multiple image region;And
The image region having a mapping relation with described representativeness mark is counted, with the number of described object obtained in described image,
Wherein, the representative mark calculating described image comprises: based on the pixel value of each pixel in described image, calculate the local extremal region of described image, using described local extremal region as described representativeness mark.
The method of attached note 12. according to attached note 11, comprises further:
The image region with mapping relation will be marked as the representative area of described object with described representativeness, and optionally non-representative area except described representative area in described multiple image region is merged into described representative area, to obtain the exterior contour figure of described object.
The method of attached note 13. according to attached note 12, wherein, optionally non-representative area except described representative area in described multiple image region is merged into described representative area to comprise: when one in described non-representative area and described representativeness mark is overlapped mutually, described non-representative area is merged into and a described representative area with mapping relation in described representativeness mark.
The method of attached note 14. according to attached note 12, wherein, optionally non-representative area except described representative area in described multiple image region is merged into described representative area to comprise: when multiple in described non-representative area and described representativeness mark are overlapped mutually, described multiple representativeness mark is chosen the representative mark that the area overlapped mutually with described non-representative area is maximum, and described non-representative area is merged into the representative area with a described representative mark with mapping relation.
The method of attached note 15. according to attached note 12 or 13, wherein, optionally non-representative area except described representative area in described multiple image region is merged into described representative area to comprise: when described non-representative area is covered by the minimum enclosed rectangle of in described representative area, or when the ratio of the area of the area that the minimum enclosed rectangle of in described representative area and described non-representative area overlap mutually and described non-representative area is greater than predetermined threshold, described non-representative area is merged in described representative area described one.
The method of attached note 16. according to attached note 11, wherein, each mapping in described representativeness mark is comprised in described multiple image region: when described representativeness mark is covered by described multiple image region, described representativeness mark is mapped to described in described multiple image region.
The method of attached note 17. according to attached note 11 or 16, wherein, each mapping in described representativeness mark is comprised in described multiple image region: when described representativeness mark overlaps mutually with the more than one image region in described multiple image region, described representativeness mark is mapped to an image region in described more than one image region.
The method of attached note 18. according to attached note 17, wherein, maps described representativeness mark and comprises to an image region in described more than one image region:
When there is the image region not covering other representative mark in the middle of described more than one image region, in the middle of described image region not covering other representativeness mark, choose the image region that the area overlapped mutually with described representativeness mark is maximum, mark the image region with mapping relation as with described representativeness;And
When in the middle of described more than one image region, each image region is covered with other representative mark, in the middle of described more than one image region, choose the image region that the area overlapped mutually with described representativeness mark is maximum, mark the image region with mapping relation as with described representativeness.
The method of attached note 19. according to attached note 11, wherein, comprised further before described image carrying out over-segmentation and calculates the representative mark of described image and described image is carried out smoothing processing.
Attached note 20. 1 kinds of machinable mediums, it carries the program product comprising the machine readable instructions code being stored therein, wherein, described instruction code is when being read by computer and perform, it is possible to described computer is performed according to method described in any one in attached note 11-19.
Claims (10)
1. an image processing apparatus, comprising:
Acquiring unit, it comprises the image of at least one object for obtaining;
Cutting unit, it is for carrying out over-segmentation to described image, to obtain multiple image region;
Calculating unit, it is for calculating the representative mark of described image;
Map unit, it is for mapping each in described representativeness mark in described multiple image region, and at least two representative marks in wherein said representativeness mark map to the same image region in described multiple image region; And
Counting unit, it is for counting the image region having a mapping relation with described representativeness mark, with the number of described object obtained in described image,
Wherein, described calculating unit, based on the pixel value of each pixel in described image, calculates the local extremal region of described image, using described local extremal region as described representativeness mark.
2. device according to claim 1, comprises further:
Merge cells, it for marking the image region with mapping relation as the representative area of described object with described representativeness, optionally non-representative area except described representative area in described multiple image region is merged into described representative area, to obtain the exterior contour figure of described object.
3. device according to claim 2, wherein, when in described non-representative area and described representativeness mark one is overlapped mutually, described non-representative area is merged into and the described representative area with mapping relation in described representativeness mark by described merge cells.
4. device according to claim 2, wherein, described non-representative area and described representativeness mark in multiple overlap mutually time, described merge cells chooses a maximum representative mark of the area overlapped mutually with described non-representative area in described multiple representativeness mark, and described non-representative area is merged into the representative area with a described representative mark with mapping relation.
5. device according to Claims 2 or 3, wherein, when described non-representative area is covered by the minimum enclosed rectangle of in described representative area, or when the ratio of the area of the area that the minimum enclosed rectangle of in described representative area and described non-representative area overlap mutually and described non-representative area is greater than predetermined threshold, described merge cells described that described non-representative area is merged in described representative area.
6. device according to claim 1, wherein, when described representativeness mark is covered by described multiple image region, described representativeness mark is mapped to described in described multiple image region by described map unit.
7. device according to claim 1 or 6, wherein, when described representativeness mark overlaps mutually with the more than one image region in described multiple image region, described representativeness mark is mapped to an image region in described more than one image region by described map unit.
8. device according to claim 7, wherein,
When there is the image region not covering other representative mark in the middle of described more than one image region, described map unit chooses the maximum image region of the area overlapped mutually with described representativeness mark in the middle of described image region not covering other representativeness mark, marks the image region with mapping relation as with described representativeness; And
When in the middle of described more than one image region, each image region is covered with other representative mark, described map unit chooses the maximum image region of the area overlapped mutually with described representativeness mark in the middle of described more than one image region, marks the image region with mapping relation as with described representativeness.
9. device according to claim 1, comprises processing unit further, and it for carrying out smoothing processing to described image before the representative mark that described image is carried out over-segmentation by described cutting unit and described calculating unit calculates described image.
10. an image processing method, comprising:
Obtain the image comprising at least one object;
Described image is carried out over-segmentation, to obtain multiple image region;
Calculate the representative mark of described image;
Mapping each in described representativeness mark in described multiple image region, at least two representative marks in wherein said representativeness mark map to the same image region in described multiple image region; And
The image region having a mapping relation with described representativeness mark is counted, with the number of described object obtained in described image,
Wherein, the representative mark calculating described image comprises: based on the pixel value of each pixel in described image, calculate the local extremal region of described image, using described local extremal region as described representativeness mark.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410648544.9A CN105654457A (en) | 2014-11-14 | 2014-11-14 | Device and method for processing image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410648544.9A CN105654457A (en) | 2014-11-14 | 2014-11-14 | Device and method for processing image |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN105654457A true CN105654457A (en) | 2016-06-08 |
Family
ID=56478901
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410648544.9A Pending CN105654457A (en) | 2014-11-14 | 2014-11-14 | Device and method for processing image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105654457A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112233114A (en) * | 2020-12-10 | 2021-01-15 | 江苏金恒信息科技股份有限公司 | Method and device for judging repeated counting of steel billets |
| CN113506301A (en) * | 2021-07-27 | 2021-10-15 | 四川九洲电器集团有限责任公司 | Tooth image segmentation method and device |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5637874A (en) * | 1994-12-20 | 1997-06-10 | Biosensor Laboratories Co., Ltd. | Apparatus and method for measuring chemiluminescence |
| WO2008080403A8 (en) * | 2006-11-16 | 2008-11-06 | Visiopharm As | Feature-based registration of sectional images |
| CN102676633A (en) * | 2012-03-08 | 2012-09-19 | 天津大学 | Method for automatically counting bacterial colonies |
| CN103020639A (en) * | 2012-11-27 | 2013-04-03 | 河海大学 | Method for automatically identifying and counting white blood cells |
| CN103561626A (en) * | 2011-06-03 | 2014-02-05 | 奥林巴斯株式会社 | Fluorescence observation device and fluorescence observation method |
| CN103760140A (en) * | 2014-01-07 | 2014-04-30 | 李雁 | Tumor metastasis unit counting method based on quantum dot spectrum analysis and image parsing |
-
2014
- 2014-11-14 CN CN201410648544.9A patent/CN105654457A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5637874A (en) * | 1994-12-20 | 1997-06-10 | Biosensor Laboratories Co., Ltd. | Apparatus and method for measuring chemiluminescence |
| WO2008080403A8 (en) * | 2006-11-16 | 2008-11-06 | Visiopharm As | Feature-based registration of sectional images |
| CN103561626A (en) * | 2011-06-03 | 2014-02-05 | 奥林巴斯株式会社 | Fluorescence observation device and fluorescence observation method |
| CN102676633A (en) * | 2012-03-08 | 2012-09-19 | 天津大学 | Method for automatically counting bacterial colonies |
| CN103020639A (en) * | 2012-11-27 | 2013-04-03 | 河海大学 | Method for automatically identifying and counting white blood cells |
| CN103760140A (en) * | 2014-01-07 | 2014-04-30 | 李雁 | Tumor metastasis unit counting method based on quantum dot spectrum analysis and image parsing |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112233114A (en) * | 2020-12-10 | 2021-01-15 | 江苏金恒信息科技股份有限公司 | Method and device for judging repeated counting of steel billets |
| CN113506301A (en) * | 2021-07-27 | 2021-10-15 | 四川九洲电器集团有限责任公司 | Tooth image segmentation method and device |
| CN113506301B (en) * | 2021-07-27 | 2024-02-23 | 四川九洲电器集团有限责任公司 | Tooth image segmentation method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20210183097A1 (en) | Spare Part Identification Using a Locally Learned 3D Landmark Database | |
| CN110874594A (en) | Human body surface damage detection method based on semantic segmentation network and related equipment | |
| CN110705583B (en) | Cell detection model training method, device, computer equipment and storage medium | |
| CN108399386A (en) | Information extracting method in pie chart and device | |
| CN110175502A (en) | A kind of backbone Cobb angle measuring method, device, readable storage medium storing program for executing and terminal device | |
| EP2908287B1 (en) | Image segmentation device, image segmentation method, and depth map generating method | |
| US11403766B2 (en) | Method and device for labeling point of interest | |
| CN105719243B (en) | Image processing apparatus and method | |
| CN112699775A (en) | Certificate identification method, device and equipment based on deep learning and storage medium | |
| CN110189341B (en) | Image segmentation model training method, image segmentation method and device | |
| CN112949766A (en) | Target area detection model training method, system, device and medium | |
| CN107977624A (en) | A kind of semantic segmentation method, apparatus and system | |
| CN110910445B (en) | Object size detection method, device, detection equipment and storage medium | |
| CN111144215A (en) | Image processing method, image processing device, electronic equipment and storage medium | |
| CN112102929A (en) | Medical image labeling method and device, storage medium and electronic equipment | |
| CN113870196B (en) | Image processing method, device, equipment and medium based on anchor point cut graph | |
| CN112966687B (en) | Image segmentation model training method and device and communication equipment | |
| CN110874170A (en) | Image area correction method, image segmentation method and device | |
| CN111368860A (en) | Relocation method and terminal equipment | |
| CN105654457A (en) | Device and method for processing image | |
| CN113762109A (en) | Training method of character positioning model and character positioning method | |
| CN116137061B (en) | Training method and device for quantity statistical model, electronic equipment and storage medium | |
| CN103837135A (en) | Workpiece detecting method and system | |
| CN115797533A (en) | Model edge tracing method, device, equipment and storage medium | |
| CN115984559A (en) | Intelligent sample selection method and related device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20160608 |