WO2022233223A1 - 图像拼接方法、装置、设备及介质 - Google Patents

图像拼接方法、装置、设备及介质 Download PDF

Info

Publication number
WO2022233223A1
WO2022233223A1 PCT/CN2022/086739 CN2022086739W WO2022233223A1 WO 2022233223 A1 WO2022233223 A1 WO 2022233223A1 CN 2022086739 W CN2022086739 W CN 2022086739W WO 2022233223 A1 WO2022233223 A1 WO 2022233223A1
Authority
WO
WIPO (PCT)
Prior art keywords
image
target
template
spliced
preset
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.)
Ceased
Application number
PCT/CN2022/086739
Other languages
English (en)
French (fr)
Inventor
何茜
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing Zitiao Network Technology Co Ltd
Original Assignee
Beijing Zitiao Network Technology Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Beijing Zitiao Network Technology Co Ltd filed Critical Beijing Zitiao Network Technology Co Ltd
Priority to US18/558,048 priority Critical patent/US12555197B2/en
Publication of WO2022233223A1 publication Critical patent/WO2022233223A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/50—Image enhancement or restoration using two or more images, e.g. averaging or subtraction
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/20—Image preprocessing
    • G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
    • G06V10/75—Organisation of the matching processes, e.g. simultaneous or sequential comparisons of image or video features; Coarse-fine approaches, e.g. multi-scale approaches; using context analysis; Selection of dictionaries
    • G06V10/751—Comparing pixel values or logical combinations thereof, or feature values having positional relevance, e.g. template matching
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20212—Image combination
    • G06T2207/20221—Image fusion; Image merging
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30196—Human being; Person
    • G06T2207/30201—Face

Definitions

  • the present disclosure relates to the technical field of image processing, for example, to an image stitching method, apparatus, device, and medium.
  • the present disclosure provides an image stitching method, device, device and medium.
  • the present disclosure provides an image stitching method, including:
  • the template image whose matching degree satisfies the preset matching condition is used as the target template image
  • the target image part in the image to be spliced is spliced into the target image area in the target template image to obtain the target spliced image, and the target image part and the target image area correspond to the same part of the target object.
  • an image splicing device comprising:
  • a first acquiring unit configured to acquire an image to be spliced and a plurality of template images, wherein the image to be spliced and the template image respectively include target objects;
  • a first processing unit configured to calculate the matching degree of the target object between the image to be spliced and each template image
  • the second processing unit is configured to use the template image whose matching degree satisfies the preset matching condition as the target template image;
  • the image stitching unit is configured to stitch the target image part in the image to be stitched into the target image area in the target template image to obtain the target stitched image, and the target image part and the target image area correspond to the same part of the target object.
  • an image stitching device including:
  • the processor is configured to read executable instructions from the memory and execute the executable instructions to implement the image stitching method described in the first aspect.
  • the present disclosure provides a computer-readable storage medium, where a computer program is stored in the storage medium, and when the computer program is executed by a processor, the processor enables the processor to implement the image stitching method described in the first aspect.
  • FIG. 1 is a schematic flowchart of an image stitching method according to an embodiment of the present disclosure
  • FIG. 2 is a schematic diagram of a template image provided by an embodiment of the present disclosure
  • FIG. 3 is a schematic diagram of a target stitching image provided by an embodiment of the present disclosure.
  • FIG. 4 is a schematic diagram of another target stitching image provided by an embodiment of the present disclosure.
  • FIG. 5 is a schematic diagram of yet another target mosaic image provided by an embodiment of the present disclosure.
  • FIG. 6 is a schematic flowchart of another image stitching method according to an embodiment of the present disclosure.
  • FIG. 7 is a schematic structural diagram of an image stitching device according to an embodiment of the present disclosure.
  • FIG. 8 is a schematic structural diagram of an image stitching device according to an embodiment of the present disclosure.
  • the term “including” and variations thereof are open-ended inclusions, ie, "including but not limited to”.
  • the term “based on” is “based at least in part on.”
  • the term “one embodiment” means “at least one embodiment”; the term “another embodiment” means “at least one additional embodiment”; the term “some embodiments” means “at least some embodiments”. Relevant definitions of other terms will be given in the description below.
  • Embodiments of the present disclosure provide an image stitching method, device, device, and medium that can meet a user's personalized needs for image stitching.
  • the image stitching method provided by the embodiments of the present disclosure is first described below with reference to FIGS. 1 to 6 .
  • the image stitching method may be performed by an image stitching device.
  • FIG. 1 shows a schematic flowchart of an image stitching method provided by an embodiment of the present disclosure.
  • the image stitching method may include the following steps.
  • the image splicing device may first acquire images to be spliced to be spliced, and after acquiring the images to be spliced, acquire multiple template images that can be used as image splicing templates of the images to be spliced.
  • the images to be stitched may include a target object.
  • the target object may include overall objects such as characters, animals, and cartoon characters, and the target object may also include partial objects such as face, head, head and shoulders, and upper body in the overall object, which are not limited herein.
  • the image stitching device may capture an image in real time, and use the image as the image to be stitched.
  • the image splicing device may also acquire an image selected locally by the user, and use the image as the image to be spliced.
  • the image splicing device may receive images sent by other devices, and use the images as images to be spliced.
  • the template image may also include a target object.
  • the target object may include overall objects such as characters, animals, and cartoon characters, and the target object may also include partial objects such as face, head, head and shoulders, and upper body in the overall object, which are not limited herein.
  • the template image may be an image in a template image library, which may contain multiple types of authorized images.
  • the template image may include, but is not limited to, a famous painting type image, an emoticon type image, a hand-drawn type image, a cartoon type image, a photo type image, etc., which is not limited in the present disclosure.
  • the image stitching device may acquire all template images in the template image library after acquiring the images to be stitched.
  • the image splicing device may, after acquiring the image to be spliced, detect the object type to which the target object in the image to be spliced belongs, and then acquire a template image with the target object corresponding to the object type in the template image library .
  • the image stitching device may acquire a template image corresponding to the person type in the template image library.
  • the image splicing device may acquire a template image corresponding to the face type in the template image library.
  • the image splicing device may obtain a template image corresponding to the person's face type in the template image library.
  • the image splicing device may also acquire multiple template images first, and then acquire images to be spliced for which image splicing needs to be performed, and then perform an image splicing operation based on the acquired multiple template images and images to be spliced .
  • the image splicing device may perform matching calculation between the target object in the image to be spliced and the target object in each template image respectively, and obtain the matching degree of the target object between the image to be spliced and each template image.
  • the matching degree of the target object can be used to represent the similarity between the target object in the image to be stitched and the target object in the template image.
  • the image stitching device may calculate the matching degree of the target object between the image to be stitched and each template image by using the object feature points of the target object.
  • S120 may include:
  • the matching degree of the target object between the image to be spliced and each template image is calculated.
  • the image splicing device first determines the to-be-spliced object feature points of the target objects in the to-be-spliced images, and then according to the template object feature points of the target objects in each template image and The mapping relationship between the feature points of the objects to be spliced of the target objects in the images to be spliced, the template object feature points of the target objects in each template image are respectively mapped to the images to be spliced, that is, the target objects in each template image are mapped.
  • the template object feature points of the target object in the to-be-spliced image are respectively subjected to feature point size uniformity processing, so that the template object feature points of the target object in each template image are the same as the target object in the to-be-spliced image.
  • the feature points of the objects to be spliced are aligned, and the feature points of the template objects of the target objects in each template image corresponding to the feature points of the target objects located in the images to be spliced are obtained, that is, the mapping of the target objects corresponding to each template image.
  • the feature points of the object to be spliced and the feature points of the object to be spliced of the target object in the image to be spliced and the feature points of the mapped object of the target object corresponding to each template image are respectively matched and calculated to obtain the target between the image to be spliced and each template image.
  • the matching degree of the object is respectively matched and calculated to obtain the target between the image to be spliced and each template image.
  • the image splicing device may use a pre-trained object feature point detection model to perform object feature point detection on the image to be spliced, and obtain object feature points of the target object in the image to be spliced.
  • the image stitching device may also acquire template object feature points of the target object in each template image stored in advance.
  • the template object feature points of the target object in each template image can be detected by using a pre-trained object feature point detection model.
  • the matching degree may include an average pixel difference between the pixels to which the feature points of the mapping object belong and the pixels to which the feature points of the object to be spliced belong. That is, the average pixel difference may be the difference between the pixel value of the pixel to which each feature point of the mapping object belongs in the image to be spliced and the pixel value of the corresponding feature point in the feature point of the object to be spliced in the image to be spliced. Average value of pixel value difference.
  • the image splicing device can extract the pixel value of the pixel to which each feature point of the target object corresponding to the template image belongs to the pixel in the image to be spliced and the target in the image to be spliced.
  • the pixel value of each feature point in the feature points of the object to be spliced is the pixel value of the pixel to which it belongs in the image to be spliced, then calculates the pixel value difference between the two pixel values corresponding to each feature point, and finally calculates the difference between all pixel values The average of the values to get the average pixel difference.
  • the matching degree may also include the average distance between the pixels to which the feature points of the mapping object belong and the pixels to which the feature points of the object to be spliced belong. That is, the average distance can be the pixel position between the pixel position of each feature point in the feature point of the mapping object in the image to be spliced and the pixel position of the corresponding feature point in the feature point of the object to be spliced in the image to be spliced. average distance.
  • the image splicing device can determine the pixel position of each feature point in the feature points of the mapped object of the target object corresponding to the template image, the pixel position of the pixel in the image to be spliced, and the target in the image to be spliced
  • the pixel position of each feature point in the feature points of the object to be spliced is the pixel position of the pixel in the image to be spliced, then calculate the pixel distance between the two pixel positions corresponding to each feature point, and finally calculate the average value of all pixel distances , to get the average distance.
  • the matching degree of the target object between the image to be spliced and each template image can be calculated more accurately by using the object feature points.
  • the image stitching device may calculate the matching degree of the target object between the image to be stitched and each template image by using the object image corresponding to the target object.
  • S120 may include:
  • the object images of the target object in each template image are respectively mapped to the images to be spliced to obtain a mapping image of each template image for the target object;
  • the image similarity between the object image of the target object in the image to be spliced and the mapped image of each template image for the target object is taken as the matching degree of the target object between the image to be spliced and each template image.
  • the image splicing device may, according to the mapping relationship between the object image of the target object in each template image and the object image of the target object in the image to be spliced
  • the object image of the object in each template image is respectively mapped to the image to be spliced, that is, the object image of the target object in each template image is subjected to image size uniformity processing to the image to be spliced, so that the target object is in each template image.
  • the object image in the image is aligned with the object image of the target object in the image to be spliced, and the mapping image located in the image to be spliced corresponding to the object image of each target object in each template image is obtained, that is, each template image is aimed at the target image.
  • the mapping image of the object and then the image similarity calculation is performed on the object image of the target object in the image to be spliced and the mapped image of each template image for the target object, to obtain the target object between the image to be spliced and each template image. suitability.
  • the image splicing device may use a pre-trained object detection model to perform object detection on the image to be spliced, and obtain an object image of the target object in the image to be spliced.
  • the image stitching device may also acquire pre-stored object images of the target object in each template image.
  • the object image of the target object in each template image can be obtained by using a pre-trained object detection model to perform object detection.
  • the image splicing device may filter the template images according to the matching degree, and filter out the matching degree satisfying the matching degree in the template image.
  • the template image with matching conditions is preset, and the filtered template image is used as the target template image, that is, the source image used for image splicing of the images to be spliced.
  • the preset matching condition may include any one of the following: the average pixel difference is more The smallest among the template images; the average pixel difference is less than or equal to the preset pixel difference threshold.
  • the preset matching condition may include that the average pixel difference is the smallest among the multiple template images, that is, the target template image may be the template image corresponding to the smallest average pixel difference among the multiple average pixel differences.
  • the image stitching device can sort the average pixel differences in ascending order. , and then take the sorted first average pixel difference as the minimum average pixel difference, and take the template image corresponding to the minimum average pixel difference as the target template image.
  • the preset matching condition may include that the average pixel difference is less than or equal to the preset pixel difference threshold, that is, the target template image may be an average pixel whose at least one of the multiple average pixel differences is less than or equal to the preset pixel difference threshold The template image corresponding to the difference.
  • the image stitching device can compare the average pixel differences with the preset pixel difference threshold respectively. , to find at least one average pixel difference less than or equal to the preset pixel difference threshold, and then use the template image corresponding to the found at least one average pixel difference as the target template image.
  • the preset pixel difference threshold may be any value preset as required, which is not limited herein.
  • the preset matching condition may include any of the following: The smallest in the template image; the average distance is less than or equal to the preset distance threshold.
  • the preset matching condition may include that the average distance is the smallest among the plurality of template images, that is, the target template image may be the template image corresponding to the smallest average distance among the plurality of average distances.
  • the image stitching device can sort the average distances in ascending order, and then The first average distance after sorting is used as the minimum average distance, and the template image corresponding to the minimum average distance is used as the target template image.
  • the preset matching condition may include that the average distance is less than or equal to a preset distance threshold, that is, the target template image may be a template corresponding to an average distance where at least one of the multiple average distances is less than or equal to the preset distance threshold image.
  • the image stitching device can compare the average distances with the preset distance thresholds to find to at least one average distance less than or equal to the preset distance threshold, and then use the template image corresponding to the at least one average distance found as the target template image.
  • the preset distance threshold may be any value preset as required, which is not limited herein.
  • the preset matching condition may include any of the following One: the image similarity is the smallest among multiple template images; the image similarity is less than or equal to the preset similarity threshold.
  • the preset matching condition may include that the image similarity is the smallest among the multiple template images, that is, the target template image may be the template image corresponding to the smallest image similarity among the multiple image similarities.
  • the image splicing device may, after calculating the image similarity between the object image of the target object in the image to be spliced and the mapped image of each template image for the target object, compare multiple images in ascending order. Then, the first image similarity after sorting is used as the minimum image similarity, and the template image corresponding to the minimum image similarity is used as the target template image.
  • the preset matching condition may include that the image similarity is less than or equal to the preset similarity threshold, that is, the target template image may be similar to images with at least one of the multiple image similarities less than or equal to the preset similarity threshold The template image corresponding to the degree.
  • the image splicing device may calculate the image similarity between the object image of the target object in the image to be spliced and the mapped image of each template image for the target object, and then compare the multiple image similarities with the preset similarity respectively.
  • the thresholds are compared to find at least one image similarity less than or equal to the preset similarity threshold, and then the template image corresponding to the found at least one image similarity is used as the target template image.
  • the preset similarity threshold may be any value preset as required, which is not limited herein.
  • the image stitching device may stitch the target image part in the to-be-spliced image into the target image area in the target template image, so as to replace the target image part in the to-be-spliced image with the target image part.
  • the original image content in the target image area in the target template image is used to obtain the target spliced image, so that the target spliced image can simultaneously have at least part of the characteristics of the target object in the image to be spliced.
  • the target image part may include at least part of the image content of the object image of the target object in the image to be spliced, that is, the target image part may include image content corresponding to at least part of the target object in the image to be spliced.
  • the target image portion and the target image area correspond to the same portion of the target object.
  • the target image part is the image content corresponding to a certain part of the target object in the image to be spliced
  • the target image area is the image area corresponding to the same part of the target object in the target template image.
  • a target template image whose matching degree with the target object between the images to be spliced satisfies a preset matching condition can be selected for the images to be spliced, so as to be based on the relationship between the template image and the image to be spliced.
  • the matching degree of the target object realizes the automatic matching of the source image used for image stitching, and then stitches the target image part corresponding to at least part of the target object in the image to be stitched to the same part of the target object in the target template image.
  • the target stitched image is obtained, so that the target stitched image can meet the user's personalized needs for image stitching, and the user experience is improved.
  • the target object in order to meet the personalized needs of the user, may include a face.
  • the face may be the face of any object.
  • Objects may include, but are not limited to, characters, animals, cartoon characters, and the like.
  • the matching degree of the target object between the image to be spliced and each template image is the matching degree of the face between the image to be spliced and each template image.
  • the image splicing device can use facial feature points of the face to calculate the matching degree of the target object between the image to be spliced and each template image.
  • the calculation of the matching degree is as described above, and will not be repeated here. .
  • the image splicing device may use the face image corresponding to the face to calculate the matching degree of the face between the image to be spliced and each template image. Repeat.
  • the image stitching device can stitch the target image part in the face image in the image to be stitched into the target image area in the image area to which the face image in the target template image belongs to obtain the target stitched image.
  • the target image portion and the portion of the target object corresponding to the target image region may be the target facial part in the face.
  • the target facial area may include at least one of the following: eyebrows, eyes, ears, nose, and mouth.
  • the target image portion may include a specified size and specified size corresponding to the target template image in the image to be stitched An image portion of a shape that contains the overall target face part.
  • the target image area may be an entire image area with the same specified size and specified shape in the target template image.
  • the above-mentioned designated size and designated shape can be preset according to requirements, which are not limited herein.
  • the specified shape may include at least one of a rectangle, a trapezoid, an ellipse, and the like.
  • the target image part in the image to be spliced can be an image part containing two eyes with a specified size and a specified shape
  • the target image area can be the target template image with the specified size and shape.
  • the size and the specified shape of an entire image area containing two eyes will be described below by taking Figures 2 and 3 as examples.
  • FIG. 2 shows a schematic diagram of a template image provided by an embodiment of the present disclosure.
  • FIG. 3 shows a schematic diagram of a target mosaic image provided by an embodiment of the present disclosure.
  • the character in the template image 201 has a template face 202 .
  • the character in the target mosaic image 301 has a mosaic face 302 , and a rectangular overall eye image 303 is mosaicked in the mosaic face 302 , and the overall eye image 303 combines the template shown in FIG. 2
  • the two eyes in the human face 202 , the image content between the two eyes, and the image content around the two eyes are entirely covered, thereby generating the target mosaic image 301 .
  • the target image part when the target face part includes eyebrows, eyes, and ears, the target image part may also be a target face with a specified size and a specified shape corresponding to the target template image and including splits part of the two images.
  • the target image areas may be two separate image areas having the same specified size and specified shape.
  • the above-mentioned designated size and designated shape can be preset according to requirements, which are not limited herein.
  • the specified shape may include at least one of a rectangle, a trapezoid, an ellipse, and the like.
  • the target image part in the image to be spliced can also be two image parts containing an eye with a specified size and a specified shape, and the target image area can be the specified size and shape.
  • the designated shape includes two separate image regions of one eye respectively, which are described below by taking FIG. 2 and FIG. 4 as examples.
  • FIG. 4 shows a schematic diagram of another target mosaic image provided by an embodiment of the present disclosure.
  • the person in the target stitched image 401 has a stitched face 402, and the stitched face 402 is stitched with two rectangular split eye images 303, and each split eye image 303 is a One eye and the image content around one eye in the template face 202 shown in 2 are covered, thereby generating the target mosaic image 401 .
  • the target image portion may further include at least two image portions each containing an entire face part with a specified size and a specified shape corresponding to the target template image.
  • the target image area may include at least two whole-part areas having the same designated size and designated shape as each image part, respectively.
  • the above-mentioned designated size and designated shape can be preset according to requirements, which are not limited herein.
  • the specified shape may include at least one of a rectangle, a trapezoid, an ellipse, and the like.
  • the target image part in the image to be spliced may include an image part with a specified size and a specified shape including two eyes, and an image part with a specified size and a specified shape including two eyes.
  • a part image of the mouth, the target image area may include a whole part area containing two eyes with the same specified size and specified shape as the image part to which the eyes belong and with the same specified size and specified shape as the image part to which the mouth belongs.
  • the shape includes an overall part area of the mouth, which will be described below by taking FIG. 2 and FIG. 5 as examples.
  • FIG. 5 shows a schematic diagram of yet another target mosaic image provided by an embodiment of the present disclosure.
  • the character in the target mosaic image 501 has a mosaic face 502, and a rectangular overall eye image 503 and a rectangular overall mouth image 504 are mosaicked in the mosaic face 502.
  • the overall eye image 503 The two eyes, the image content between the two eyes, and the image content around the two eyes in the template face 202 shown in FIG. 2 are covered as a whole, and the whole mouth image 504 is shown in FIG.
  • the mouth in the template face 202 and the image content around the mouth are covered as a whole, thereby generating the target mosaic image 501 .
  • a target template image whose matching degree with its face satisfies a preset matching condition can be selected for the image to be spliced from among the multiple template images, so as to be based on the template image and the face of the image to be spliced. to achieve automatic matching of the source image, and then splicing the target image part corresponding to the target face part in the image to be spliced to the target image area to which the target face part belongs in the selected target template image to obtain the target spliced image , so that the target mosaic image can meet the personalized needs of users and improve the user experience.
  • the mapping relationship between the template object feature points of the target object in the template image and the to-be-spliced object feature points of the target object in the to-be-spliced image may include template object feature points to to-be-spliced objects The optimal affine transformation matrix of the feature points.
  • mapping the template object feature points of the target object in the template image to the image to be spliced, and obtaining the mapped object feature points of the target object corresponding to the template image may include:
  • the feature points of the template object are respectively mapped to the image to be spliced to obtain the feature points of the mapped object.
  • the image splicing device can perform affine transformation processing on each template object feature point to the feature point of the object to be spliced, so as to calculate the optimal affine transformation matrix from each template object feature point to the feature point of the object to be spliced, and then Multiply each template object feature point with the corresponding optimal affine transformation matrix to obtain the mapping object feature point corresponding to each template object feature point.
  • the image splicing device may traverse multiple template images, and for each template image, the template object feature points of the target object in the template image may be affine to the object feature points to be spliced of the target object in the image to be spliced. Transformation processing to calculate the optimal affine transformation matrix from the template object feature points of the target object in the template image to the feature points of the target object to be stitched in the template image, and then the target object in the template image.
  • Each feature point in the feature points of the template object is multiplied by the corresponding optimal affine transformation matrix to realize the mapping of each feature point in the feature points of the template object of the target object in the template image to the image to be spliced, Then, the mapped face feature points corresponding to the template object feature points of the target object in the template image are obtained.
  • the image splicing device may use the mapping relationship between the template object feature points of the target object in the template image and the to-be-spliced object feature points of the target object in the to-be-spliced image as the target object in each template image
  • the mapping relationship between the object image of the target object and the object image of the target object in the image to be spliced The mapping relationship with the object image of the target object in the image to be spliced.
  • the image stitching device calculates the template object feature points of the target object in the template image to the to-be-stitched object feature points of the target object in the to-be-stitched image. After the optimal affine transformation matrix of The mapping of each pixel of the template image to the image to be spliced, and then the mapping image of the template image for the target object is obtained.
  • the optimal affine transformation matrix from the feature points of the template object to the feature points of the objects to be spliced can be used to realize the transformation between the template object feature points of the target object in the template image and the target object in the template image.
  • the mapping of the object image to the image to be spliced improves the accuracy of the mapping result, thereby improving the accuracy of the acquired target template image.
  • the image splicing device may use the splicing mask image corresponding to the target template image to implement image splicing between the image to be spliced and the target template image.
  • S140 shown in FIG. 1 may include:
  • the target image part in the mapped target template image is stitched into the target image area in the target template image to obtain the target stitched image.
  • the image splicing device can map the image to be spliced onto the target template image according to the mapping relationship between the target object in the image to be spliced and the target object in the template image, that is, the image to be spliced is converted to the target template image by the size of the image.
  • the homogenization process makes the image to be spliced and the target template image aligned to obtain the mapped target template image, and then uses the stitching mask image corresponding to the target template image to fuse the mapped target template image and the target template image, so as to obtain the target template image after stitching.
  • the mask image realizes that the target image part in the mapped target template image is spliced into the target image area in the target template image, and the target spliced image is obtained.
  • the mapping relationship between the target object in the image to be spliced and the target object in the template image may include the optimal target template object feature point of the target object in the target template image to the feature point of the object to be spliced in the image to be spliced Affine transformation inverse matrix.
  • mapping the image to be spliced onto the target template image, and obtaining the mapped target template image may include:
  • the image to be spliced is mapped onto the target template image to obtain the mapped target template image.
  • the image stitching device may obtain an optimal affine transformation matrix from the template object feature points of the target object in the target template image to the feature points of the target object to be stitched in the image to be stitched, and calculate the optimal affine transformation matrix of the The inverse matrix is obtained to obtain the optimal affine transformation inverse matrix of the target template object feature point of the target object in the target template image to the feature point of the object to be spliced in the image to be spliced, and then each pixel of the image to be spliced is compared with the optimal affine transformation matrix.
  • the affine transformation inverse matrix is multiplied to realize the mapping of each pixel of the image to be spliced to the target template image, and then the mapped target template image corresponding to the image to be spliced is obtained, so that the target object in the mapped target template image is The size and position are the same as those of the target object in the target template image.
  • the image stitching device may obtain a stitching mask image corresponding to the target template image from a plurality of stitching mask images stored in advance.
  • the stitching mask image corresponding to the target template image can be used to extract the target image part located in the target image area in the target template image and the original image part outside the target image area in the target template image in the target template image after mapping , and the extracted target image part and the original image part are stitched into the target stitched image.
  • the image stitching method may further include:
  • a stitched mask image is generated.
  • the image stitching device can firstly use the pre-trained object feature point detection model to detect the object feature points of each template image respectively, obtain the template object feature points of the target object in each template image, and then determine the template object feature points according to the preset area.
  • the method based on the feature points of the template object, determine the target image area to which the target part of the target object in each template image belongs, and then generate the stitching mask image corresponding to each template image according to the target image area, so that the stitching mask image can be It is used to extract the target image part to which the target part of the target object in the mapped target template image belongs and the original image part outside the target image area to which the target part of the target object in the target template image belongs, and the extracted target image part And the original image part is stitched into the target stitched image.
  • determining the target image area according to the target template object feature points by the image stitching device may include: determining, among the template object feature points of the target object in the template image, the feature points corresponding to the target part of the target object, and then calculating the feature points that can surround the determined target object. The minimum image area with the specified shape of the feature points, and finally the minimum image area is enlarged by a predetermined ratio to obtain the target image area.
  • the image stitching device can determine the feature points corresponding to the target face part from the facial feature points in the template image, and then Calculate the minimum image area with the specified shape that can surround the determined feature point, and finally expand the minimum image area by a predetermined ratio to obtain the target image area.
  • the image stitching method may also include:
  • the multiple template images that meet the preset template conditions are filtered.
  • the image splicing device may first acquire a plurality of preset images, and then perform image detection on each preset image to determine whether each preset image conforms to the preset template conditions, and then filter out those conforming to the preset Multiple template images of template conditions, and store the filtered template images in the template image library.
  • the preset template conditions may include at least one of the following: the image proportion of the target object is greater than or equal to a preset ratio threshold; the object angle of the target object is less than or equal to the preset angle threshold; the object size of the target object is greater than or equal to Preset size threshold.
  • the object angle may include the attitude angle of the target object, such as Euler angle, which is not limited herein.
  • an image with a target object and the target object meets the requirements can be used as a template image, and further, after the image to be spliced and the template image are spliced, the characteristics of the spliced target object can be made more obvious , which makes it easier for users to view the stitching effect, which not only improves the quality of image stitching, but also improves the user experience.
  • the preset template conditions may include at least one of the following: the image proportion of the face is greater than or equal to the preset ratio threshold; the face angle is less than or equal to the preset angle threshold; the face size Greater than or equal to the preset size threshold.
  • the image stitching device can use the face detection model obtained by pre-training to perform face detection on the preset image, and obtain the image in the preset image. face image, and calculate the ratio of the face image to the preset image, the ratio is the image proportion of the face, and then the image stitching device can filter out the image proportion of the face is greater than or equal to the preset ratio threshold. set the image, and use the filtered preset image as the template image.
  • the preset ratio threshold can be preset as required, which is not limited here.
  • the image stitching device can use the pre-trained face pose detection model to perform facial pose detection on the preset image, and obtain the The face posture angle is the face angle, and then the image stitching device can filter out a preset image whose face angle is less than or equal to the preset angle threshold, and use the filtered preset image as a template image.
  • the preset angle threshold may be preset as required, and the facial posture angle may include the facial Euler angle, which is not limited herein.
  • the image stitching device can use the face detection model obtained by pre-training to perform face detection on the preset image, and obtain the face in the preset image. image, and calculate the size of the face image, which is the face size, and then the image stitching device can filter out the preset image whose face size is greater than or equal to the preset size threshold, and use the filtered preset image as a template image.
  • the preset size threshold may be preset as required, which is not limited herein.
  • an image with a face that meets the requirements can be used as a template image, and further, after the image to be spliced and the template image are spliced, the facial features of the spliced face can be made more obvious , which makes it easier for users to view the stitching effect, which not only improves the quality of image stitching, but also improves the user experience.
  • the image stitching method shown in FIG. 6 is also provided.
  • FIG. 6 shows a schematic flowchart of another image stitching method provided by an embodiment of the present disclosure.
  • the image stitching method may include the following steps.
  • the image to be spliced and the template image respectively include target objects.
  • the target image part and the target image area correspond to the same part of the target object.
  • S610-S640 are similar to S110-S140 shown in FIG. 1 , and details are not described here.
  • the number of target mosaic images is multiple.
  • the target object includes the face, the target image part, and the part of the target object corresponding to the target image area is the target face part in the face.
  • the image stitching device can respectively stitch the target image parts corresponding to each target face part in the images to be stitched into the corresponding target template images.
  • the target image area to which the target face part belongs and the combined image of the target image parts corresponding to various target face parts in the image to be spliced are spliced into the combined area of the target image area to which the corresponding target face part belongs in the target template image , to obtain multiple target stitched images.
  • the image stitching device may stitch the target image parts in the images to be stitched to the target image areas in each target template image respectively, so as to obtain multiple targets Stitched images.
  • S650 Generate a target video according to a preset video template and a plurality of target mosaic images.
  • a preset video template may be obtained, and the preset video template is used to automatically perform video editing on the multiple target spliced images to obtain the target video.
  • a single image to be spliced can be automatically matched to at least one target template image with a similar target object, and the image to be spliced and the at least one target template image can be spliced in various forms. , to obtain multiple target stitched images, and then generate a target video composed of multiple target stitched graphics, so as to facilitate the user to create interesting videos and improve the user's experience.
  • An embodiment of the present disclosure further provides an image stitching device for implementing the above-mentioned image stitching method, which will be described below with reference to FIG. 7 .
  • the image stitching apparatus may be performed by an image stitching device.
  • FIG. 7 shows a schematic structural diagram of an image stitching apparatus provided by an embodiment of the present disclosure.
  • the image stitching apparatus 700 may include a first acquisition unit 710 , a first processing unit 720 , a second processing unit 730 and an image stitching unit 740 .
  • the first acquiring unit 710 may be configured to acquire an image to be spliced and a plurality of template images, wherein the image to be spliced and the template images respectively include target objects.
  • the first processing unit 720 may be configured to calculate the matching degree of the target object between the image to be stitched and each template image.
  • the second processing unit 730 may be configured to use the template image whose matching degree satisfies the preset matching condition as the target template image.
  • the image stitching unit 740 can be configured to stitch the target image part in the image to be stitched into the target image area in the target template image to obtain the target stitched image, and the target image part and the target image area correspond to the same part of the target object.
  • a target template image whose matching degree with the target object between the images to be spliced satisfies a preset matching condition can be selected for the images to be spliced, so as to be based on the relationship between the template image and the image to be spliced.
  • the matching degree of the target object realizes the automatic matching of the source image used for image stitching, and then stitches the target image part corresponding to at least part of the target object in the image to be stitched to the same part of the target object in the target template image.
  • the target stitched image is obtained, so that the target stitched image can meet the user's personalized needs for image stitching, and the user experience is improved.
  • the first processing unit 720 may include a first sub-processing unit, a second sub-processing unit, and a third sub-processing unit.
  • the first sub-processing unit may be configured to determine the feature points of the object to be spliced of the target object in the image to be spliced.
  • the second sub-processing unit may be configured to, for each template image, map template object feature points of the target object in the template image to the image to be spliced to obtain the mapped object feature points of the target object corresponding to the template image.
  • the third sub-processing unit may be configured to calculate the matching degree of the target object between the image to be spliced and each template image according to the feature points of the object to be spliced and the feature points of each mapped object.
  • the second sub-processing unit may be configured to calculate an optimal affine transformation matrix from the feature points of the template object to the feature points of the object to be spliced; Map to the image to be spliced to obtain the feature points of the mapping object.
  • the matching degree includes an average pixel difference between the pixel to which the feature point of the mapping object belongs and the pixel to which the feature point of the object to be spliced belongs.
  • the matching condition may include any one of the following: the average pixel difference is the smallest among the plurality of template images; the average pixel difference is less than or equal to a preset pixel difference threshold.
  • the matching degree includes an average distance between the pixels to which the feature points of the mapping object belong and the pixels to which the feature points of the object to be spliced belong;
  • the matching condition may include any one of the following: the average distance is the smallest among the plurality of template images; the average distance is less than or equal to a preset distance threshold.
  • the image stitching unit 740 may include a fourth sub-processing unit and a fifth sub-processing unit.
  • the fourth sub-processing unit may be configured to map the image to be spliced onto the target template image to obtain the mapped target template image.
  • the fifth sub-processing unit may be configured to, based on the stitching mask image corresponding to the target template image, stitch the target image part in the mapped target template image into the target image area in the target template image to obtain the target stitched image.
  • the fourth sub-processing unit may be configured to calculate the optimal affine affine of the target template object feature points of the target object in the target template image to the to-be-spliced object feature points of the target object in the to-be-spliced image Transform inverse matrix; map the image to be spliced onto the target template image according to the optimal affine transformation inverse matrix to obtain the mapped target template image.
  • the image stitching apparatus 700 may further include a third processing unit and a fourth processing unit.
  • the third processing unit may be configured to determine the target image area according to the feature points of the target template object.
  • the fourth processing unit may be configured to generate a mosaic mask image according to the target image area.
  • the image stitching apparatus 700 may further include a second acquiring unit and a fifth processing unit.
  • the second acquiring unit may be configured to acquire multiple preset images before acquiring the image to be spliced and the multiple template images.
  • the fifth processing unit may be configured to select, among the plurality of preset images, a plurality of template images that meet the preset template conditions.
  • the preset template condition may include at least one of the following: the image ratio of the target object is greater than or equal to the preset ratio threshold; the object angle of the target object is less than or equal to the preset angle threshold; the object size of the target object is greater than or equal to the preset angle threshold or equal to the preset size threshold.
  • the target object includes a face.
  • the number of the target mosaic images may be multiple.
  • the image splicing apparatus 700 may further include a video generating unit, and the video generating unit may be configured to generate a target video according to a preset video template and a plurality of target splicing images after obtaining the target splicing image.
  • the image stitching apparatus 700 shown in FIG. 7 can perform multiple steps in the method embodiments shown in FIGS. 1 to 6 , and implement multiple steps in the method embodiments shown in FIGS. 1 to 6 . The process and effect will not be repeated here.
  • Embodiments of the present disclosure also provide an image stitching device, the image stitching device may include a processor and a memory, and the memory may be used to store executable instructions.
  • the processor may be configured to read executable instructions from the memory, and execute the executable instructions to implement the image stitching method in the above embodiment.
  • FIG. 8 shows a schematic structural diagram of an image stitching device provided by an embodiment of the present disclosure. Referring next to FIG. 8 , it shows a schematic structural diagram of an image stitching apparatus 800 suitable for implementing an embodiment of the present disclosure.
  • image stitching device 800 shown in FIG. 8 is only an example, and should not impose any limitations on the functions and scope of use of the embodiments of the present disclosure.
  • the image stitching apparatus 800 may include a processing device (eg, a central processing unit, a graphics processor, etc.) 801 , which may be loaded from a storage device 808 according to a program stored in a read-only memory (ROM) 802 or from a storage device 808 .
  • Programs in random access memory (RAM) 803 perform various appropriate actions and processes.
  • RAM 803 In the RAM 803, various programs and data required for the operation of the image stitching device 800 are also stored.
  • the processing device 801, the ROM 802, and the RAM 803 are connected to each other through a bus 804.
  • An input/output (I/O) interface 805 is also connected to bus 804 .
  • the following devices may be connected to the I/O interface 805: input devices 806 including, for example, a touch screen, touchpad, keyboard, mouse, camera, microphone, accelerometer, gyroscope, etc.; including, for example, a liquid crystal display (LCD), speakers, vibration An output device 807 of a computer, etc.; a storage device 808 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 809.
  • the communication means 809 may allow the image stitching device 800 to communicate wirelessly or by wire with other devices to exchange data.
  • FIG. 8 shows the image stitching apparatus 800 having various means, it should be understood that not all of the illustrated means are required to be implemented or available. More or fewer devices may alternatively be implemented or provided.
  • Embodiments of the present disclosure further provide a computer-readable storage medium, where a computer program is stored in the storage medium, and when the computer program is executed by the processor, the processor enables the processor to implement the image stitching method in the foregoing embodiment.
  • embodiments of the present disclosure include a computer program product comprising a computer program carried on a non-transitory computer readable medium, the computer program containing program code for performing the method illustrated in the flowchart.
  • the computer program may be downloaded and installed from the network via the communication device 809, or from the storage device 808, or from the ROM 802.
  • the processing device 801 the above-mentioned functions defined in the image stitching method of the embodiment of the present disclosure are executed.
  • the computer-readable medium mentioned above in the present disclosure may be a computer-readable signal medium or a computer-readable storage medium, or any combination of the above two.
  • the computer readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus or device, or a combination of any of the above. More specific examples of computer readable storage media may include, but are not limited to, electrical connections with one or more wires, portable computer disks, hard disks, random access memory (RAM), read only memory (ROM), erasable Programmable read only memory (EPROM or flash memory), fiber optics, portable compact disk read only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.
  • a computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.
  • a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave with computer-readable program code embodied thereon. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the foregoing.
  • a computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device .
  • Program code embodied on a computer readable medium may be transmitted using any suitable medium including, but not limited to, electrical wire, optical fiber cable, RF (radio frequency), etc., or any suitable combination of the foregoing.
  • clients, servers can communicate using any currently known or future developed network protocol, such as HTTP, and can be interconnected with any form or medium of digital data communication (eg, a communication network).
  • a communication network examples include local area networks (“LAN”), wide area networks (“WAN”), the Internet (eg, the Internet), and peer-to-peer networks (eg, ad hoc peer-to-peer networks), as well as any currently known or future development network of.
  • the above-mentioned computer-readable medium may be included in the above-mentioned image splicing apparatus; or may exist alone without being assembled into the image splicing apparatus.
  • the above-mentioned computer-readable medium carries one or more programs, and when the above-mentioned one or more programs are executed by the image splicing device, the image splicing device is made to execute:
  • the image to be spliced and a plurality of template images wherein the image to be spliced and the template image respectively include target objects; calculate the matching degree of the target object between the image to be spliced and each template image;
  • the template image is used as the target template image; the target image part in the image to be spliced is spliced into the target image area in the target template image to obtain the target spliced image, and the target image part and the target image area correspond to the same part of the target object.
  • computer program code for performing operations of the present disclosure may be written in one or more programming languages, including but not limited to object-oriented programming languages—such as Java, Smalltalk, C++, and also conventional procedural programming languages - such as the "C" language or similar programming languages.
  • the program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer, and the like.
  • the remote computer may be connected to the user's computer through any kind of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (eg, using an Internet service provider through Internet connection).
  • LAN local area network
  • WAN wide area network
  • Internet service provider e.g., using an Internet service provider through Internet connection
  • each block in the flowchart or block diagrams may represent a module, segment, or portion of code that contains one or more logical functions for implementing the specified functions executable instructions.
  • the functions noted in the blocks may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved.
  • each block of the block diagrams and/or flowchart illustrations, and combinations of blocks in the block diagrams and/or flowchart illustrations can be implemented in dedicated hardware-based systems that perform the specified functions or operations , or can be implemented in a combination of dedicated hardware and computer instructions.
  • the units involved in the embodiments of the present disclosure may be implemented in a software manner, and may also be implemented in a hardware manner. Among them, the name of the unit does not constitute a limitation of the unit itself under certain circumstances.
  • exemplary types of hardware logic components include: Field Programmable Gate Arrays (FPGAs), Application Specific Integrated Circuits (ASICs), Application Specific Standard Products (ASSPs), Systems on Chips (SOCs), Complex Programmable Logical Devices (CPLDs) and more.
  • FPGAs Field Programmable Gate Arrays
  • ASICs Application Specific Integrated Circuits
  • ASSPs Application Specific Standard Products
  • SOCs Systems on Chips
  • CPLDs Complex Programmable Logical Devices
  • a machine-readable medium may be a tangible medium that may contain or store a program for use by or in connection with the instruction execution system, apparatus or device.
  • the machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium.
  • Machine-readable media may include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or devices, or any suitable combination of the foregoing.
  • machine-readable storage media would include one or more wire-based electrical connections, portable computer disks, hard disks, random access memory (RAM), read only memory (ROM), erasable programmable read only memory (EPROM or flash memory), fiber optics, compact disk read only memory (CD-ROM), optical storage, magnetic storage, or any suitable combination of the foregoing.
  • RAM random access memory
  • ROM read only memory
  • EPROM or flash memory erasable programmable read only memory
  • CD-ROM compact disk read only memory
  • magnetic storage or any suitable combination of the foregoing.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Multimedia (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Medical Informatics (AREA)
  • Software Systems (AREA)
  • Databases & Information Systems (AREA)
  • Computing Systems (AREA)
  • Artificial Intelligence (AREA)
  • Health & Medical Sciences (AREA)
  • Image Analysis (AREA)

Abstract

本公开涉及一种图像拼接方法、装置、设备及介质。其中，图像拼接方法包括：获取待拼接图像和多个模板图像，其中，待拼接图像和模板图像分别包括目标对象；计算待拼接图像与每个模板图像之间的目标对象的匹配度；将匹配度满足预设匹配条件的模板图像作为目标模板图像；将待拼接图像中的目标图像部分拼接至目标模板图像中的目标图像区域内，得到目标拼接图像，目标图像部分和目标图像区域对应目标对象的相同部分。

Description

图像拼接方法、装置、设备及介质
本申请要求在2021年5月7日提交中国专利局、申请号为202110496619.6的中国专利申请的优先权，该申请的全部内容通过引用结合在本申请中。
技术领域
本公开涉及图像处理技术领域，例如涉及一种图像拼接方法、装置、设备及介质。
背景技术
随着计算机技术和移动通信技术的迅速发展，越来越多的应用程序走进了人们的生活，极大地丰富了人们的日常生活。这些应用程序提供了多种交互功能，受到人们的喜爱。
然而，用户还期望应用程序能够提供更多满足用户个性化需求、提升用户体验的交互功能。
发明内容
本公开提供了一种图像拼接方法、装置、设备及介质。
第一方面，本公开提供了一种图像拼接方法，包括：
获取待拼接图像和多个模板图像，其中，待拼接图像和模板图像分别包括目标对象；
计算待拼接图像与每个模板图像之间的目标对象的匹配度；
将匹配度满足预设匹配条件的模板图像作为目标模板图像；
将待拼接图像中的目标图像部分拼接至目标模板图像中的目标图像区域内，得到目标拼接图像，目标图像部分和目标图像区域对应目标对象的相同部分。
第二方面，本公开提供了一种图像拼接装置，包括：
第一获取单元，配置为获取待拼接图像和多个模板图像，其中，待拼接图像和模板图像分别包括目标对象；
第一处理单元，配置为计算待拼接图像与每个模板图像之间的目标对象的匹配度；
第二处理单元，配置为将匹配度满足预设匹配条件的模板图像作为目标模板图像；
图像拼接单元，配置为将待拼接图像中的目标图像部分拼接至目标模板图像中的目标图像区域内，得到目标拼接图像，目标图像部分和目标图像区域对应目标对象的相同部分。
第三方面，本公开提供了一种图像拼接设备，包括：
处理器；
存储器，用于存储可执行指令；
其中，处理器用于从存储器中读取可执行指令，并执行可执行指令以实现第一方面所述的图像拼接方法。
第四方面，本公开提供了一种计算机可读存储介质，该存储介质存储有计算机程序，当计算机程序被处理器执行时，使得处理器实现第一方面所述的图像拼接方法。
附图说明
贯穿附图中，相同或相似的附图标记表示相同或相似的元素。应当理解附图是示意性的，原件和元素不一定按照比例绘制。
图1为本公开实施例提供的一种图像拼接方法的流程示意图；
图2为本公开实施例提供的一种模板图像的示意图；
图3为本公开实施例提供的一种目标拼接图像的示意图；
图4为本公开实施例提供的另一种目标拼接图像的示意图；
图5为本公开实施例提供的又一种目标拼接图像的示意图；
图6为本公开实施例提供的另一种图像拼接方法的流程示意图；
图7为本公开实施例提供的一种图像拼接装置的结构示意图；
图8为本公开实施例提供的一种图像拼接设备的结构示意图。
具体实施方式
下面将参照附图更详细地描述本公开的实施例。虽然附图中显示了本公开的某些实施例，然而应当理解的是，本公开可以通过多种形式来实现，而且不应该被解释为限于这里阐述的实施例，相反提供这些实施例是为了更加透彻和完整地理解本公开。应当理解的是，本公开的附图及实施例仅用于示例性作用，并非用于限制本公开的保护范围。
应当理解，本公开的方法实施方式中记载的多个步骤可以按照不同的顺序执行，和/或并行执行。此外，方法实施方式可以包括附加的步骤和/或省略执行示出的步骤。本公开的范围在此方面不受限制。
本文使用的术语“包括”及其变形是开放性包括，即“包括但不限于”。术语“基于”是“至少部分地基于”。术语“一个实施例”表示“至少一个实施例”；术语“另一实施例”表示“至少一个另外的实施例”；术语“一些实施例”表示“至少一些实施例”。其他术语的相关定义将在下文描述中给出。
需要注意，本公开中提及的“第一”、“第二”等概念仅用于对不同的装置、模块或单元进行区分，并非用于限定这些装置、模块或单元所执行的功能的顺序或者相互依存关系。
需要注意，本公开中提及的“一个”、“多个”的修饰是示意性而非限制性的，本领域技术人员应当理解，除非在上下文另有明确指出，否则应该理解为“一个或多个”。
本公开实施方式中的多个装置之间所交互的消息或者信息的名称仅用于说明性的目的，而并不是用于对这些消息或信息的范围进行限制。
本公开实施例提供了一种能够满足用户对图像拼接的个性化需求的图像拼接方法、装置、设备及介质。
下面首先结合图1至图6对本公开实施例提供的图像拼接方法进行说明。在本公开实施例中，该图像拼接方法可以由图像拼接设备执行。
图1示出了本公开实施例提供的一种图像拼接方法的流程示意图。
如图1所示，该图像拼接方法可以包括如下步骤。
S110、获取待拼接图像和多个模板图像。
在本公开实施例中，图像拼接设备可以首先获取需要进行图像拼接的待拼接图像，并且在获取到待拼接图像之后，再获取可作为该待拼接图像的图像拼接模板的多个模板图像。
在本公开实施例中，待拼接图像可以包括目标对象。
例如，目标对象可以包括人物、动物、卡通形象等整体对象，目标对象也可以包括整体对象中的脸部、头部、头肩部、上半身等局部对象，在此不作限制。
在一些实施例中，图像拼接设备可以实时拍摄图像，并将该图像作为待拼接图像。
在另一些实施例中，图像拼接设备也可以获取用户在本地所选择的图像，并将该图像作为待拼接图像。
在又一些实施例中，图像拼接设备可以接收其他设备所发送的图像，并将该图像作为待拼接图像。
在本公开实施例中，模板图像也可以包括目标对象。
例如，目标对象可以包括人物、动物、卡通形象等整体对象，目标对象也可以包括整体对象中的脸部、头部、头肩部、上半身等局部对象，在此不作限制。
例如，模板图像可以为模板图像库中的图像，该模板图像库中可以收录有获得授权的多种类型的图像。
例如，模板图像可以包括但不限于名画类型的图像、表情包类型的图像、手绘类型的图像、漫画类型的图像、照片类型的图像等，本公开不对此进行限制。
在一些实施例中，图像拼接设备可以在获取待拼接图像之后，获取模板图像库中的全部模板图像。
在另一些实施例中，图像拼接设备可以在获取待拼接图像之后，检测该待拼接图像中的目标对象所属的对象类型，进而在模板图像库中获取具有该对象类型对应的目标对象的模板图像。
例如，在待拼接图像中的目标对象所属的对象类型为人物类型的情况下，图像拼接设备可以在模板图像库中获取人物类型对应的模板图像。再例如，在待拼接图像中的目标对象所属的对象类型为脸部类型的情况下，图像拼接设备可以在模板图像库中获取脸部类型对应的模板图像。又例如，在待拼接图像中的目标对象所属的对象类型为人物脸部类型的情况下，图像拼接设备可以在模板图像库中获取人物脸部类型对应的模板图像。
此外，在又一些实施例中，图像拼接设备还可以先获取多个模板图像，再获取需要进行图像拼接的待拼接图像，进而基于获取到的多个模板图像和待拼接图像，进行图像拼接操作。
S120、计算待拼接图像与每个模板图像之间的目标对象的匹配度。
在本公开实施例中，图像拼接设备可以将待拼接图像中的目标对象分别与每个模板图像中的目标对象进行匹配计算，得到待拼接图像与每个模板图像之间的目标对象的匹配度。
例如，目标对象的匹配度可以用于表征待拼接图像中的目标对象与模板图像中的目标对象之间的相似程度。
在一些实施例中，图像拼接设备可以利用目标对象的对象特征点计算待拼接图像与每个模板图像之间的目标对象的匹配度。
在这些实施例中，S120可以包括：
确定待拼接图像中的目标对象的待拼接对象特征点；
针对每个模板图像，将模板图像中的目标对象的模板对象特征点映射至待拼接图像上，得到模板图像对应的目标对象的映射对象特征点；
根据待拼接对象特征点和每个映射对象特征点，计算待拼接图像与每个模板图像之间的 目标对象的匹配度。
例如，图像拼接设备在获取到待拼接图像和多个模板图像之后，首先确定待拼接图像中的目标对象的待拼接对象特征点，然后按照每个模板图像中的目标对象的模板对象特征点与待拼接图像中的目标对象的待拼接对象特征点之间的映射关系，将每个模板图像中的目标对象的模板对象特征点分别映射至待拼接图像上，即将每个模板图像中的目标对象的模板对象特征点分别向待拼接图像中的目标对象的待拼接对象特征点进行特征点尺寸的均一化处理，使得每个模板图像中的目标对象的模板对象特征点与待拼接图像中的目标对象的待拼接对象特征点对齐，得到每个模板图像中的目标对象的模板对象特征点对应的位于待拼接图像内的目标对象的映射对象特征点，即每个模板图像对应的目标对象的映射对象特征点，进而对待拼接图像中的目标对象的待拼接对象特征点和每个模板图像对应的目标对象的映射对象特征点分别进行匹配计算，得到待拼接图像与每个模板图像之间的目标对象的匹配度。
例如，图像拼接设备可以利用预先训练得到的对象特征点检测模型对待拼接图像进行对象特征点检测，得到待拼接图像中的目标对象的待拼接对象特征点。
例如，图像拼接设备还可以获取预先存储的每个模板图像中的目标对象的模板对象特征点。其中，每个模板图像中的目标对象的模板对象特征点均可以利用预先训练得到的对象特征点检测模型进行检测得到。
在这些实施例中，匹配度可以包括映射对象特征点所属像素和待拼接对象特征点所属像素的平均像素差。即平均像素差可以为映射对象特征点中的每个特征点在待拼接图像中所属像素的像素值与待拼接对象特征点中的对应特征点在待拼接图像中所属像素的像素值之间的像素值差值的平均值。
例如，针对待拼接图像和一个模板图像，图像拼接设备可以提取模板图像对应的目标对象的映射对象特征点中的每个特征点在待拼接图像中所属像素的像素值以及待拼接图像中的目标对象的待拼接对象特征点中的每个特征点在待拼接图像中所属像素的像素值，然后计算每个特征点对应的两个像素值之间的像素值差值，最后计算所有像素值差值的平均值，得到平均像素差。
在这些实施例中，匹配度也可以包括映射对象特征点所属像素与待拼接对象特征点所属像素的平均距离。即平均距离可以为映射对象特征点中的每个特征点在待拼接图像中所属像素的像素位置与待拼接对象特征点中的对应特征点在待拼接图像中所属像素的像素位置之间的像素距离的平均值。
例如，针对待拼接图像和一个模板图像，图像拼接设备可以确定模板图像对应的目标对象的映射对象特征点中的每个特征点在待拼接图像中所属像素的像素位置以及待拼接图像中的目标对象的待拼接对象特征点中的每个特征点在待拼接图像中所属像素的像素位置，然后计算每个特征点对应的两个像素位置之间的像素距离，最后计算所有像素距离的平均值，得到平均距离。
由此，在本公开实施例中，可以通过对象特征点较为精确地计算待拼接图像与每个模板图像之间的目标对象的匹配度。
在另一些实施例中，图像拼接设备可以利用目标对象对应的对象图像计算待拼接图像与每个模板图像之间的目标对象的匹配度。
在这些实施例中，S120可以包括：
将目标对象在每个模板图像中的对象图像分别映射至待拼接图像上，得到每个模板图像针对目标对象的映射图像；
将目标对象在待拼接图像中的对象图像与每个模板图像针对目标对象的映射图像之间的图像相似度，作为待拼接图像与每个模板图像之间的目标对象的匹配度。
例如，图像拼接设备在获取到待拼接图像和多个模板图像之后，可以按照目标对象在每个模板图像中的对象图像与目标对象在待拼接图像中的对象图像之间的映射关系，将目标对象在每个模板图像中的对象图像分别映射至待拼接图像上，即将目标对象在每个模板图像中的对象图像分别向待拼接图像进行图像尺寸的均一化处理，使得目标对象在每个模板图像中的对象图像与目标对象在待拼接图像中的对象图像对齐，得到每个目标对象在每个模板图像中的对象图像对应的位于待拼接图像内的映射图像，即每个模板图像针对目标对象的映射图像，进而对目标对象在待拼接图像中的对象图像和每个模板图像针对目标对象的映射图像分别进行图像相似度计算，得到待拼接图像与每个模板图像之间的目标对象的匹配度。
例如，图像拼接设备可以利用预先训练得到的对象检测模型对待拼接图像进行对象检测，得到目标对象在待拼接图像中的对象图像。
例如，图像拼接设备还可以获取预先存储的目标对象在每个模板图像中的对象图像。其中，目标对象在每个模板图像中的对象图像均可以利用预先训练得到的对象检测模型进行对象检测得到。
S130、将匹配度满足预设匹配条件的模板图像作为目标模板图像。
在本公开实施例中，图像拼接设备可以在获得待拼接图像与每个模板图像之间的目标对象的匹配度之后，根据匹配度对模板图像进行筛选，并在模板图像中筛选出匹配度满足预设匹配条件的模板图像，以将筛选出的模板图像作为目标模板图像，即用于对待拼接图像进行图像拼接的源图像。
在一些实施例中，在匹配度包括映射对象特征点所属像素和待拼接对象特征点所属像素的平均像素差的情况下，预设匹配条件可以包括下列中的任一项：平均像素差在多个模板图像中最小；平均像素差小于或等于预设像素差阈值。
在一个示例中，预设匹配条件可以包括平均像素差在多个模板图像中最小，即目标模板图像可以为多个平均像素差中的最小平均像素差所对应的模板图像。
例如，图像拼接设备可以在计算得到每个模板图像对应的映射对象特征点所属像素和待拼接对象特征点所属像素的平均像素差之后，按照由小到大的顺序对多个平均像素差进行排序，进而将排序后的首个平均像素差作为最小平均像素差，并将该最小平均像素差所对应的模板图像作为目标模板图像。
在另一个示例中，预设匹配条件可以包括平均像素差小于或等于预设像素差阈值，即目标模板图像可以为多个平均像素差中的至少一个小于或等于预设像素差阈值的平均像素差所对应的模板图像。
例如，图像拼接设备可以在计算得到每个模板图像对应的映射对象特征点所属像素和待拼接对象特征点所属像素的平均像素差之后，将多个平均像素差分别与预设像素差阈值进行比较，以查找到至少一个小于或等于预设像素差阈值的平均像素差，进而将查找到的至少一个平均像素差所对应的模板图像作为目标模板图像。
其中，预设像素差阈值可以为根据需要预先设置的任意值，在此不作限制。
在另一些实施例中，在匹配度包括映射对象特征点所属像素与待拼接对象特征点所属像素的平均距离的情况下，预设匹配条件可以包括下列中的任一项：平均距离在多个模板图像中最小；平均距离小于或等于预设距离阈值。
在一个示例中，预设匹配条件可以包括平均距离在多个模板图像中最小，即目标模板图像可以为多个平均距离中的最小平均距离所对应的模板图像。
例如，图像拼接设备可以在计算得到每个模板图像对应的映射对象特征点所属像素与待拼接对象特征点所属像素的平均距离之后，按照由小到大的顺序对多个平均距离进行排序，进而将排序后的首个平均距离作为最小平均距离，并将该最小平均距离所对应的模板图像作为目标模板图像。
在另一个示例中，预设匹配条件可以包括平均距离小于或等于预设距离阈值，即目标模板图像可以为多个平均距离中的至少一个小于或等于预设距离阈值的平均距离所对应的模板图像。
例如，图像拼接设备可以在计算得到每个模板图像对应的映射对象特征点所属像素与待拼接对象特征点所属像素的平均距离之后，将多个平均距离分别与预设距离阈值进行比较，以查找到至少一个小于或等于预设距离阈值的平均距离，进而将查找到的至少一个平均距离所对应的模板图像作为目标模板图像。
其中，预设距离阈值可以为根据需要预先设置的任意值，在此不作限制。
在又一些实施例中，在匹配度包括目标对象在待拼接图像中的对象图像与模板图像针对目标对象的映射图像之间的图像相似度的情况下，预设匹配条件可以包括下列中的任一项：图像相似度在多个模板图像中最小；图像相似度小于或等于预设相似度阈值。
在一个示例中，预设匹配条件可以包括图像相似度在多个模板图像中最小，即目标模板图像可以为多个图像相似度中的最小图像相似度所对应的模板图像。
例如，图像拼接设备可以在计算得到目标对象在待拼接图像中的对象图像与每个模板图像针对目标对象的映射图像之间的图像相似度之后，按照由小到大的顺序对多个图像相似度进行排序，进而将排序后的首个图像相似度作为最小图像相似度，并将该最小图像相似度所对应的模板图像作为目标模板图像。
在另一个示例中，预设匹配条件可以包括图像相似度小于或等于预设相似度阈值，即目标模板图像可以为多个图像相似度中的至少一个小于或等于预设相似度阈值的图像相似度所对应的模板图像。
例如，图像拼接设备可以在计算得到目标对象在待拼接图像中的对象图像与每个模板图像针对目标对象的映射图像之间的图像相似度之后，将多个图像相似度分别与预设相似度阈值进行比较，以查找到至少一个小于或等于预设相似度阈值的图像相似度，进而将查找到的至少一个图像相似度所对应的模板图像作为目标模板图像。
其中，预设相似度阈值可以为根据需要预先设置的任意值，在此不作限制。
S140、将待拼接图像中的目标图像部分拼接至目标模板图像中的目标图像区域内，得到目标拼接图像。
在本公开实施例中，图像拼接设备可以在确定目标模板图像之后，将待拼接图像中的目标图像部分拼接至目标模板图像中的目标图像区域内，以利用待拼接图像中的目标图像部分替换目标模板图像中的目标图像区域内的原始图像内容，得到目标拼接图像，使得该目标拼 接图像可以同时具有目标对象在待拼接图像中的至少部分特征。
其中，目标图像部分可以包括目标对象在待拼接图像中的对象图像的至少部分图像内容，即目标图像部分可以包括目标对象的至少部分在待拼接图像中所对应的图像内容。
例如，目标图像部分和目标图像区域对应目标对象的相同部分。例如，目标图像部分如果为目标对象的某部分在待拼接图像中所对应的图像内容，则目标图像区域则为目标对象的相同部分在目标模板图像中所对应的图像区域。
在本公开实施例中，能够在多个模板图像中，为待拼接图像选取与其之间的目标对象的匹配度满足预设匹配条件的目标模板图像，以基于模板图像与待拼接图像之间的目标对象的匹配度，实现对用于进行图像拼接的源图像的自动匹配，进而将待拼接图像中的目标对象的至少部分对应的目标图像部分拼接至目标模板图像中的目标对象的相同部分对应的目标图像区域内，得到目标拼接图像，使得目标拼接图像可以满足用户对图像拼接的个性化需求，提升了用户的体验。
在本公开另一种实施方式中，为了满足用户的个性化需求，目标对象可以包括脸部。
其中，脸部可以为任意对象的脸部。对象可以包括但不限于人物、动物、卡通形象等。
例如，待拼接图像与每个模板图像之间的目标对象的匹配度即为待拼接图像与每个模板图像之间的脸部的匹配度。
在一些实施例中，图像拼接设备可以利用脸部的脸部特征点计算待拼接图像与每个模板图像之间的目标对象的匹配度，匹配度的计算如上文所述，在此不做赘述。
在另一些实施例中，图像拼接设备可以利用脸部对应的脸部图像计算待拼接图像与每个模板图像之间的脸部的匹配度，匹配度的计算如上文所述，在此不做赘述。
例如，图像拼接设备可以将待拼接图像中的脸部图像中的目标图像部分拼接至目标模板图像中的脸部图像所属图像区域中的目标图像区域内，得到目标拼接图像。
在这些实施例中，目标图像部分和目标图像区域所对应的目标对象的部分可以为脸部中的目标脸部部位。目标脸部部位可以包括下列中的至少一种：眉部、眼部、耳部、鼻部和口部。
在一些实施例中，在目标脸部部位包括眉部、眼部、耳部、鼻部或口部的情况下，目标图像部分可以包括待拼接图像中的具有目标模板图像对应的指定尺寸和指定形状的包含整体的目标脸部部位的一个图像部分。
此时，目标图像区域可以为目标模板图像中具有相同的指定尺寸和指定形状的一个整体图像区域。
其中，上述的指定尺寸和指定形状均可以根据需要预先设置，在此不作限制。指定形状可以包括矩形、梯形、椭圆形等中的至少一种。
以目标脸部部位包括眼部为例，待拼接图像中的目标图像部分可以为具有指定尺寸和指定形状的包含有两个眼睛的一个图像部分，目标图像区域可以为目标模板图像中具有该指定尺寸和该指定形状的包含有两个眼睛的一个整体图像区域，下面以图2和图3为例进行说明。
图2示出了本公开实施例提供的一种模板图像的示意图。图3示出了本公开实施例提供的一种目标拼接图像的示意图。
如图2所示，该模板图像201中的人物具有模板人脸202。如图3所示，该目标拼接图像301中的人物具有拼接人脸302，该拼接人脸302中拼接有矩形的整体眼部图像303，该整 体眼部图像303将图2中所示的模板人脸202中的两个眼睛、两个眼睛之间的图像内容及两个眼睛周边的图像内容进行了整体覆盖，从而生成了目标拼接图像301。
在另一些实施例中，在目标脸部部位包括眉部、眼部和耳部的情况下，目标图像部分也可以为具有目标模板图像对应的指定尺寸和指定形状的包含分体的目标脸部部位的两个图像部分。此时，目标图像区域可以为具有相同的指定尺寸和指定形状的两个分体图像区域。
其中，上述的指定尺寸和指定形状均可以根据需要预先设置，在此不作限制。指定形状可以包括矩形、梯形、椭圆形等中的至少一种。
继续以目标脸部部位包括眼部为例，待拼接图像中的目标图像部分也可以为具有指定尺寸和指定形状的分别包含一个眼睛的两个图像部分，目标图像区域可以为具有该指定尺寸和该指定形状的分别包含有一个眼睛的两个分体图像区域，下面以图2和图4为例进行说明。
图4示出了本公开实施例提供的另一种目标拼接图像的示意图。
如图4所示，该目标拼接图像401中的人物具有拼接人脸402，该拼接人脸402中拼接有两个矩形的分体眼部图像303，每个分体眼部图像303分别将图2中所示的模板人脸202中的一个眼睛及一个眼睛周边的图像内容进行了覆盖，从而生成了目标拼接图像401。
例如，目标图像部分还可以包括具有目标模板图像对应的指定尺寸和指定形状的分别包含一个整体脸部部位的至少两个图像部分。此时，目标图像区域可以包括分别具有与每个图像部分相同的指定尺寸和指定形状的至少两个整体部位区域。
其中，上述的指定尺寸和指定形状均可以根据需要预先设置，在此不作限制。指定形状可以包括矩形、梯形、椭圆形等中的至少一种。
以目标脸部部位包括眼部和嘴部为例，待拼接图像中的目标图像部分可以包括具有指定尺寸和指定形状的包含有两个眼睛的一个图像部分以及具有指定尺寸和指定形状的包含有嘴巴的一个部位图像，目标图像区域可以包括具有与眼睛所属的图像部分相同的指定尺寸和指定形状的包含有两个眼睛的一个整体部位区域和具有与嘴巴所属的图像部分相同的指定尺寸和指定形状的包含有嘴巴的一个整体部位区域，下面以图2和图5为例进行说明。
图5示出了本公开实施例提供的又一种目标拼接图像的示意图。
如图5所示，该目标拼接图像501中的人物具有拼接人脸502，该拼接人脸502中拼接有矩形的整体眼部图像503和矩形的整体嘴部图像504，该整体眼部图像503将图2中所示的模板人脸202中的两个眼睛、两个眼睛之间的图像内容及两个眼睛周边的图像内容进行了整体覆盖，该整体嘴部图像504将图2中所示的模板人脸202中的嘴巴及嘴巴周边的图像内容进行了整体覆盖，从而生成了目标拼接图像501。
由此，在本公开实施例中，能够在多个模板图像中，为待拼接图像选取与其脸部的匹配度满足预设匹配条件的目标模板图像，以基于模板图像与待拼接图像的脸部的匹配度，实现对源图像的自动匹配，进而将待拼接图像中目标脸部部位对应的目标图像部分拼接至所选取的目标模板图像中目标脸部部位所属的目标图像区域，得到目标拼接图像，使得目标拼接图像可以满足用户的个性化需求，提升了用户的体验。在本公开又一种实施方式中，模板图像中的目标对象的模板对象特征点与待拼接图像中的目标对象的待拼接对象特征点之间的映射关系可以包括模板对象特征点到待拼接对象特征点的最优仿射变换矩阵。
在这些实施例中，将模板图像中的目标对象的模板对象特征点映射至待拼接图像上，得到模板图像对应的目标对象的映射对象特征点可以包括：
计算模板对象特征点到待拼接对象特征点的最优仿射变换矩阵；
根据最优仿射变换矩阵，将模板对象特征点分别映射至待拼接图像上，得到映射对象特征点。
例如，图像拼接设备可以对每个模板对象特征点进行向待拼接对象特征点的仿射变换处理，以计算出每个模板对象特征点到待拼接对象特征点的最优仿射变换矩阵，然后将每个模板对象特征点分别与对应的最优仿射变换矩阵相乘，得到每个模板对象特征点对应的映射对象特征点。
例如，图像拼接设备可以遍历多个模板图像，对于每个模板图像，可以对该模板图像中的目标对象的模板对象特征点进行向待拼接图像中的目标对象的待拼接对象特征点的仿射变换处理，以计算出该模板图像中的目标对象的模板对象特征点到待拼接图像中的目标对象的待拼接对象特征点的最优仿射变换矩阵，然后将该模板图像中的目标对象的模板对象特征点中的每个特征点分别与对应的最优仿射变换矩阵相乘，实现对模板图像中的目标对象的模板对象特征点中的每个特征点向待拼接图像中的映射，进而得到该模板图像中的目标对象的模板对象特征点所对应的映射脸部特征点。
在这些实施例中，图像拼接设备可以将模板图像中的目标对象的模板对象特征点与待拼接图像中的目标对象的待拼接对象特征点之间的映射关系作为目标对象在每个模板图像中的对象图像与目标对象在待拼接图像中的对象图像之间的映射关系，即将模板对象特征点到待拼接对象特征点的最优仿射变换矩阵作为目标对象在每个模板图像中的对象图像与目标对象在待拼接图像中的对象图像之间的映射关系。
例如，图像拼接设备在遍历多个模板图像的过程中，对于每个模板图像，在计算得到该模板图像中的目标对象的模板对象特征点到待拼接图像中的目标对象的待拼接对象特征点的最优仿射变换矩阵之后，可以将目标对象在该模板图像中的对象图像的每个像素分别与对应的最优仿射变换矩阵相乘，实现对目标对象在该模板图像中的对象图像的每个像素向待拼接图像中的映射，进而得到该模板图像针对目标对象的映射图像。
由此，在本公开实施例中，可以利用模板对象特征点到待拼接对象特征点的最优仿射变换矩阵实现对模板图像中的目标对象的模板对象特征点和目标对象在模板图像中的对象图像向待拼接图像的映射，提高映射结果的准确性，进而提高所获取的目标模板图像的准确性。
在本公开再一种实施方式中，图像拼接设备可以利用目标模板图像对应的拼接掩码图像实现对待拼接图像与目标模板图像的图像拼接。
在这些实施例中，图1所示的S140可以包括：
将待拼接图像映射到目标模板图像上，得到映射后的目标模板图像；
基于目标模板图像对应的拼接掩码图像，将映射后的目标模板图像中的目标图像部分拼接至目标模板图像中的目标图像区域内，得到目标拼接图像。
例如，图像拼接设备可以按照待拼接图像中的目标对象与模板图像中的目标对象之间的映射关系，将待拼接图像映射到目标模板图像上，即将待拼接图像向目标模板图像进行图像尺寸的均一化处理，使得待拼接图像与目标模板图像对齐，得到映射后的目标模板图像，进而利用目标模板图像对应的拼接掩码图像对映射后的目标模板图像和目标模板图像进行融合，以通过拼接掩码图像实现将映射后的目标模板图像中的目标图像部分拼接至目标模板图像中的目标图像区域内，得到目标拼接图像。
例如，待拼接图像中的目标对象与模板图像中的目标对象之间的映射关系可以包括目标模板图像中的目标对象的目标模板对象特征点到待拼接图像中的待拼接对象特征点的最优仿射变换逆矩阵。
其中，将待拼接图像映射到目标模板图像上，得到映射后的目标模板图像可以包括：
计算目标模板图像中的目标对象的目标模板对象特征点到待拼接图像中的目标对象的待拼接对象特征点的最优仿射变换逆矩阵；
根据最优仿射变换逆矩阵，将待拼接图像映射到目标模板图像上，得到映射后的目标模板图像。
例如，图像拼接设备可以获取目标模板图像的目标对象的模板对象特征点到待拼接图像中的目标对象的待拼接对象特征点的最优仿射变换矩阵，并且计算该最优仿射变换矩阵的逆矩阵，得到目标模板图像中的目标对象的目标模板对象特征点到待拼接图像中的待拼接对象特征点的最优仿射变换逆矩阵，然后将待拼接图像的每个像素分别与最优仿射变换逆矩阵相乘，实现对待拼接图像的每个像素向目标模板图像中的映射，进而得到待拼接图像对应的映射后的目标模板图像，使得映射后的目标模板图像中的目标对象的尺寸和位置与目标模板图像中的目标对象的尺寸和位置相同。
例如，图像拼接设备可以从预先存储的多个拼接掩码图像中，获取目标模板图像对应的拼接掩码图像。
其中，目标模板图像对应的拼接掩码图像可以用于提取映射后的目标模板图像中位于目标模板图像中的目标图像区域中的目标图像部分和目标模板图像中的目标图像区域以外的原始图像部分，并且将提取的目标图像部分和原始图像部分拼接为目标拼接图像。
在这些实施例中，在图1所示的S140之前，该图像拼接方法还可以包括：
根据目标模板对象特征点，确定目标图像区域；
根据目标图像区域，生成拼接掩码图像。
例如，图像拼接设备可以首先利用预先训练得到的对象特征点检测模型分别对每个模板图像进行对象特征点检测，得到每个模板图像中的目标对象的模板对象特征点，然后按照预设区域确定方式，基于模板对象特征点，确定每个模板图像中的目标对象的目标部分所属的目标图像区域，进而根据目标图像区域，生成每个模板图像对应的拼接掩码图像，使得拼接掩码图像可以用于提取映射后的目标模板图像中的目标对象的目标部分所属的目标图像部分和目标模板图像中的目标对象的目标部分所属的目标图像区域以外的原始图像部分，并且将提取的目标图像部分和原始图像部分拼接为目标拼接图像。
例如，图像拼接设备根据目标模板对象特征点确定目标图像区域可以包括：在模板图像中的目标对象的模板对象特征点中，确定目标对象的目标部分对应的特征点，然后计算能够包围所确定的特征点的具有指定形状的最小图像区域，最后将该最小图像区域扩大预定比例，得到目标图像区域。
以目标对象为脸部、目标对象的目标部位为脸部中的目标脸部部位为例，图像拼接设备可以在模板图像中的脸部特征点中，确定目标脸部部位对应的特征点，然后计算能够包围所确定的特征点的具有指定形状的最小图像区域，最后将该最小图像区域扩大预定比例，得到目标图像区域。
在本公开再一种实施方式中，为了提高图像拼接质量，在图1所示的S110之前，该图 像拼接方法还可以包括：
获取多个预设图像；
在多个预设图像中，筛选符合预设模板条件的多个模板图像。
在本公开实施例中，图像拼接设备可以首先获取多个预设图像，然后对每个预设图像进行图像检测，以判断每个预设图像是否符合预设模板条件，进而筛选出符合预设模板条件的多个模板图像，并将筛选出的模板图像存储至模板图像库中。
其中，预设模板条件可以包括下列中的至少一项：目标对象的图像占比大于或等于预设比值阈值；目标对象的对象角度小于或等于预设角度阈值；目标对象的对象尺寸大于或等于预设尺寸阈值。
其中，对象角度可以包括目标对象的姿态角度如欧拉角，在此不作限制。
由此，在本公开实施例中，可以将具有目标对象且目标对象符合要求的图像作为模板图像，进而在将待拼接图像与模板图像进行拼接后，可以使拼接后的目标对象的特征更明显，使用户更容易观看到拼接效果，不但能够提高图像拼接质量，还能够提升用户的体验。
以目标对象包括脸部为例，预设模板条件可以包括下列中的至少一项：脸部的图像占比大于或等于预设比值阈值；脸部角度小于或等于预设角度阈值；脸部尺寸大于或等于预设尺寸阈值。
在预设模板条件包括脸部的图像占比大于或等于预设比值阈值的情况下，图像拼接设备可以利用预先训练得到的脸部检测模型对预设图像进行脸部检测，得到预设图像中的脸部图像，并计算脸部图像与预设图像的比值，该比值即为脸部的图像占比，接着图像拼接设备可以筛选出脸部的图像占比大于或等于预设比值阈值的预设图像，并将筛选出的预设图像作为模板图像。
其中，预设比值阈值可以根据需要预先设置，在此不作限制。
在预设模板条件包括脸部角度小于或等于预设角度阈值的情况下，图像拼接设备可以利用预先训练得到的脸部姿态检测模型对预设图像进行脸部姿态检测，得到预设图像中的脸部姿态角，该脸部姿态角即为脸部角度，接着图像拼接设备可以筛选出脸部角度小于或等于预设角度阈值的预设图像，并将筛选出的预设图像作为模板图像。
其中，预设角度阈值可以根据需要预先设置，脸部姿态角可以包括脸部欧拉角，在此不作限制。
在预设模板条件包括脸部尺寸大于或等于预设尺寸阈值的情况下，图像拼接设备可以利用预先训练得到的脸部检测模型对预设图像进行脸部检测，得到预设图像中的脸部图像，并计算脸部图像的尺寸，该尺寸即为脸部尺寸，接着图像拼接设备可以筛选出脸部尺寸大于或等于预设尺寸阈值的预设图像，并将筛选出的预设图像作为模板图像。
其中，预设尺寸阈值可以根据需要预先设置，在此不作限制。
由此，在本公开实施例中，可以将具有脸部且脸部符合要求的图像作为模板图像，进而在将待拼接图像与模板图像进行拼接后，可以使拼接人脸的人脸特征更明显，使用户更容易观看到拼接效果，不但能够提高图像拼接质量，还能够提升用户的体验。
在本公开又一种实施方式中，为了提升用户的体验，还提供了图6所示的图像拼接方法。
图6示出了本公开实施例提供的另一种图像拼接方法的流程示意图。
如图6所示，该图像拼接方法可以包括如下步骤。
S610、获取待拼接图像和多个模板图像。
其中，待拼接图像和模板图像分别包括目标对象。
S620、计算待拼接图像与每个模板图像之间的目标对象的匹配度。
S630、将匹配度满足预设匹配条件的模板图像作为目标模板图像。
S640、将待拼接图像中的目标图像部分拼接至目标模板图像中的目标图像区域内，得到目标拼接图像。
其中，目标图像部分和目标图像区域对应目标对象的相同部分。
需要说明的是，S610-S640与图1所示的S110-S140相似，在此不做赘述。
在本公开实施例中，目标拼接图像的数量为多个。
下面，以目标对象包括脸部、目标图像部分和目标图像区域所对应的目标对象的部分为脸部中的目标脸部部位为例进行说明。
在一些实施例中，当目标脸部部位包括多种脸部部位的情况下，图像拼接设备可以将待拼接图像中的每种目标脸部部位对应的目标图像部分分别拼接至目标模板图像中相应目标脸部部位所属的目标图像区域以及将待拼接图像中的多种目标脸部部位对应的目标图像部分的组合图像拼接至目标模板图像中相应目标脸部部位所属的目标图像区域的组合区域中，得到多个目标拼接图像。
在另一些实施例中，当目标模板图像的数量为多个的情况下，图像拼接设备可以将待拼接图像中的目标图像部分分别拼接至每个目标模板图像中目标图像区域，得到多个目标拼接图像。
S650、根据预设视频模板和多个目标拼接图像，生成目标视频。
在本公开实施例中，在图像拼接设备得到多个目标拼接图像之后，可以获取预设视频模板，并且利用预设视频模板自动对多个目标拼接图像进行视频剪辑，得到目标视频。
由此，在本公开实施例中，可以为单个待拼接图像自动匹配到与其具有相似目标对象的至少一个目标模板图像，并将该待拼接图像与至少一个目标模板图像进行多种形式的图像拼接，得到多个目标拼接图像，进而生成由多个目标拼接图形构成的目标视频，以方便用户创作趣味视频，提升了用户的体验。
本公开实施例还提供了一种用于实现上述的图像拼接方法的图像拼接装置，下面结合图7进行说明。在本公开实施例中，该图像拼接装置可以由图像拼接设备执行。
图7示出了本公开实施例提供的一种图像拼接装置的结构示意图。
如图7所示，该图像拼接装置700可以包括第一获取单元710、第一处理单元720、第二处理单元730和图像拼接单元740。
该第一获取单元710可以配置为获取待拼接图像和多个模板图像，其中，待拼接图像和模板图像分别包括目标对象。
该第一处理单元720可以配置为计算待拼接图像与每个模板图像之间的目标对象的匹配度。
该第二处理单元730可以配置为将匹配度满足预设匹配条件的模板图像作为目标模板图像。
该图像拼接单元740可以配置为将待拼接图像中的目标图像部分拼接至目标模板图像中的目标图像区域内，得到目标拼接图像，目标图像部分和目标图像区域对应目标对象的相同 部分。
在本公开实施例中，能够在多个模板图像中，为待拼接图像选取与其之间的目标对象的匹配度满足预设匹配条件的目标模板图像，以基于模板图像与待拼接图像之间的目标对象的匹配度，实现对用于进行图像拼接的源图像的自动匹配，进而将待拼接图像中的目标对象的至少部分对应的目标图像部分拼接至目标模板图像中的目标对象的相同部分对应的目标图像区域内，得到目标拼接图像，使得目标拼接图像可以满足用户对图像拼接的个性化需求，提升了用户的体验。
在本公开一些实施例中，该第一处理单元720可以包括第一子处理单元、第二子处理单元和第三子处理单元。
该第一子处理单元可以配置为确定待拼接图像中的目标对象的待拼接对象特征点。
该第二子处理单元可以配置为针对每个模板图像，将模板图像中的目标对象的模板对象特征点映射至待拼接图像上，得到模板图像对应的目标对象的映射对象特征点。
该第三子处理单元可以配置为根据待拼接对象特征点和每个映射对象特征点，计算待拼接图像与每个模板图像之间的目标对象的匹配度。
在本公开一些实施例中，该第二子处理单元可以配置为计算模板对象特征点到待拼接对象特征点的最优仿射变换矩阵；根据最优仿射变换矩阵，将模板对象特征点分别映射至待拼接图像上，得到映射对象特征点。
在本公开一些实施例中，该匹配度包括映射对象特征点所属的像素和待拼接对象特征点所属像素的平均像素差。
相应地，该匹配条件可以包括下列中的任一项：平均像素差在多个模板图像中最小；平均像素差小于或等于预设像素差阈值。
在本公开一些实施例中，该匹配度包括映射对象特征点所属的像素与待拼接对象特征点所属像素的平均距离；
相应地，该匹配条件可以包括下列中的任一项：平均距离在多个模板图像中最小；平均距离小于或等于预设距离阈值。
在本公开一些实施例中，该图像拼接单元740可以包括第四子处理单元和第五子处理单元。
该第四子处理单元可以配置为将待拼接图像映射到目标模板图像上，得到映射后的目标模板图像。
该第五子处理单元可以配置为基于目标模板图像对应的拼接掩码图像，将映射后的目标模板图像中的目标图像部分拼接至目标模板图像中的目标图像区域内，得到目标拼接图像。
在本公开一些实施例中，该第四子处理单元可以配置为计算目标模板图像中的目标对象的目标模板对象特征点到待拼接图像中的目标对象的待拼接对象特征点的最优仿射变换逆矩阵；根据最优仿射变换逆矩阵，将待拼接图像映射到目标模板图像上，得到映射后的目标模板图像。
在本公开一些实施例中，该图像拼接装置700还可以包括第三处理单元和第四处理单元。
该第三处理单元可以配置为根据目标模板对象特征点，确定目标图像区域。
该第四处理单元可以配置为根据目标图像区域，生成拼接掩码图像。
在本公开一些实施例中，该图像拼接装置700还可以包括第二获取单元和第五处理单元。
该第二获取单元可以配置为在获取待拼接图像和多个模板图像之前，获取多个预设图像。
该第五处理单元可以配置为在多个预设图像中，筛选符合预设模板条件的多个模板图像。
相应地，该预设模板条件可以包括下列中的至少一项：目标对象的图像占比大于或等于预设比值阈值；目标对象的对象角度小于或等于预设角度阈值；目标对象的对象尺寸大于或等于预设尺寸阈值。
在本公开一些实施例中，目标对象包括脸部。
在本公开一些实施例中，该目标拼接图像的数量可以为多个。
相应地，该图像拼接装置700还可以包括视频生成单元，该视频生成单元可以配置为在得到目标拼接图像之后，根据预设视频模板和多个目标拼接图像，生成目标视频。
需要说明的是，图7所示的图像拼接装置700可以执行图1至图6所示的方法实施例中的多个步骤，并且实现图1至图6所示的方法实施例中的多个过程和效果，在此不做赘述。
本公开实施例还提供了一种图像拼接设备，该图像拼接设备可以包括处理器和存储器，存储器可以用于存储可执行指令。其中，处理器可以用于从存储器中读取可执行指令，并执行可执行指令以实现上述实施例中的图像拼接方法。
图8示出了本公开实施例提供的一种图像拼接设备的结构示意图。下面参考图8，其示出了适于用来实现本公开实施例中的图像拼接设备800的结构示意图。
需要说明的是，图8示出的图像拼接设备800仅仅是一个示例，不应对本公开实施例的功能和使用范围带来任何限制。
如图8所示，该图像拼接设备800可以包括处理装置(例如中央处理器、图形处理器等)801，其可以根据存储在只读存储器(ROM)802中的程序或者从存储装置808加载到随机访问存储器(RAM)803中的程序而执行多种适当的动作和处理。在RAM 803中，还存储有图像拼接设备800操作所需的多种程序和数据。处理装置801、ROM 802以及RAM 803通过总线804彼此相连。输入/输出(I/O)接口805也连接至总线804。
通常，以下装置可以连接至I/O接口805：包括例如触摸屏、触摸板、键盘、鼠标、摄像头、麦克风、加速度计、陀螺仪等的输入装置806；包括例如液晶显示器(LCD)、扬声器、振动器等的输出装置807；包括例如磁带、硬盘等的存储装置808；以及通信装置809。通信装置809可以允许图像拼接设备800与其他设备进行无线或有线通信以交换数据。虽然图8示出了具有多种装置的图像拼接设备800，但是应理解的是，并不要求实施或具备所有示出的装置。可以替代地实施或具备更多或更少的装置。
本公开实施例还提供了一种计算机可读存储介质，该存储介质存储有计算机程序，当计算机程序被处理器执行时，使得处理器实现上述实施例中的图像拼接方法。
根据本公开的实施例，上文参考流程图描述的过程可以被实现为计算机软件程序。例如，本公开的实施例包括一种计算机程序产品，其包括承载在非暂态计算机可读介质上的计算机程序，该计算机程序包含用于执行流程图所示的方法的程序代码。在这样的实施例中，该计算机程序可以通过通信装置809从网络上被下载和安装，或者从存储装置808被安装，或者从ROM 802被安装。在该计算机程序被处理装置801执行时，执行本公开实施例的图像拼接方法中限定的上述功能。
需要说明的是，本公开上述的计算机可读介质可以是计算机可读信号介质或者计算机可读存储介质或者是上述两者的任意组合。计算机可读存储介质例如可以是——但不限于—— 电、磁、光、电磁、红外线、或半导体的系统、装置或器件，或者任意以上的组合。计算机可读存储介质的更具体的例子可以包括但不限于：具有一个或多个导线的电连接、便携式计算机磁盘、硬盘、随机访问存储器(RAM)、只读存储器(ROM)、可擦式可编程只读存储器(EPROM或闪存)、光纤、便携式紧凑磁盘只读存储器(CD-ROM)、光存储器件、磁存储器件、或者上述的任意合适的组合。在本公开中，计算机可读存储介质可以是任何包含或存储程序的有形介质，该程序可以被指令执行系统、装置或者器件使用或者与其结合使用。而在本公开中，计算机可读信号介质可以包括在基带中或者作为载波一部分传播的数据信号，其中承载了计算机可读的程序代码。这种传播的数据信号可以采用多种形式，包括但不限于电磁信号、光信号或上述的任意合适的组合。计算机可读信号介质还可以是计算机可读存储介质以外的任何计算机可读介质，该计算机可读信号介质可以发送、传播或者传输用于由指令执行系统、装置或者器件使用或者与其结合使用的程序。计算机可读介质上包含的程序代码可以用任何适当的介质传输，包括但不限于：电线、光缆、RF(射频)等等，或者上述的任意合适的组合。
在一些实施方式中，客户端、服务器可以利用诸如HTTP之类的任何当前已知或未来研发的网络协议进行通信，并且可以与任意形式或介质的数字数据通信(例如，通信网络)互连。通信网络的示例包括局域网(“LAN”)，广域网(“WAN”)，网际网(例如，互联网)以及端对端网络(例如，ad hoc端对端网络)，以及任何当前已知或未来研发的网络。
上述计算机可读介质可以是上述图像拼接设备中所包含的；也可以是单独存在，而未装配入该图像拼接设备中。
上述计算机可读介质承载有一个或者多个程序，当上述一个或者多个程序被该图像拼接设备执行时，使得该图像拼接设备执行：
获取待拼接图像和多个模板图像，其中，待拼接图像和模板图像分别包括目标对象；计算待拼接图像与每个模板图像之间的目标对象的匹配度；将匹配度满足预设匹配条件的模板图像作为目标模板图像；将待拼接图像中的目标图像部分拼接至目标模板图像中的目标图像区域内，得到目标拼接图像，目标图像部分和目标图像区域对应目标对象的相同部分。
在本公开实施例中，可以以一种或多种程序设计语言或其组合来编写用于执行本公开的操作的计算机程序代码，上述程序设计语言包括但不限于面向对象的程序设计语言—诸如Java、Smalltalk、C++，还包括常规的过程式程序设计语言—诸如“C”语言或类似的程序设计语言。程序代码可以完全地在用户计算机上执行、部分地在用户计算机上执行、作为一个独立的软件包执行、部分在用户计算机上部分在远程计算机上执行、等。在涉及远程计算机的情形中，远程计算机可以通过任意种类的网络——包括局域网(LAN)或广域网(WAN)—连接到用户计算机，或者，可以连接到外部计算机(例如利用因特网服务提供商来通过因特网连接)。
附图中的流程图和框图，图示了按照本公开多种实施例的系统、方法和计算机程序产品的可能实现的体系架构、功能和操作。在这点上，流程图或框图中的每个方框可以代表一个模块、程序段、或代码的一部分，该模块、程序段、或代码的一部分包含一个或多个用于实现规定的逻辑功能的可执行指令。也应当注意，在有些作为替换的实现中，方框中所标注的功能也可以以不同于附图中所标注的顺序发生。例如，两个接连地表示的方框实际上可以基本并行地执行，它们有时也可以按相反的顺序执行，这依所涉及的功能而定。也要注意的是， 框图和/或流程图中的每个方框、以及框图和/或流程图中的方框的组合，可以用执行规定的功能或操作的专用的基于硬件的系统来实现，或者可以用专用硬件与计算机指令的组合来实现。
描述于本公开实施例中所涉及到的单元可以通过软件的方式实现，也可以通过硬件的方式来实现。其中，单元的名称在某种情况下并不构成对该单元本身的限定。
本文中以上描述的功能可以至少部分地由一个或多个硬件逻辑部件来执行。例如，非限制性地，可以使用的示范类型的硬件逻辑部件包括：现场可编程门阵列(FPGA)、专用集成电路(ASIC)、专用标准产品(ASSP)、片上系统(SOC)、复杂可编程逻辑设备(CPLD)等等。
在本公开的上下文中，机器可读介质可以是有形的介质，其可以包含或存储以供指令执行系统、装置或设备使用或与指令执行系统、装置或设备结合地使用的程序。机器可读介质可以是机器可读信号介质或机器可读储存介质。机器可读介质可以包括但不限于电子的、磁性的、光学的、电磁的、红外的、或半导体系统、装置或设备，或者上述内容的任何合适组合。机器可读存储介质的更具体示例会包括基于一个或多个线的电气连接、便携式计算机盘、硬盘、随机存取存储器(RAM)、只读存储器(ROM)、可擦除可编程只读存储器(EPROM或快闪存储器)、光纤、便捷式紧凑盘只读存储器(CD-ROM)、光学储存设备、磁储存设备、或上述内容的任何合适组合。
以上描述仅为本公开的示例实施例以及对所运用技术原理的说明。本领域技术人员应当理解，本公开中所涉及的公开范围，并不限于上述技术特征的特定组合而成的技术方案，同时也应涵盖在不脱离上述公开构思的情况下，由上述技术特征或其等同特征进行任意组合而形成的其它技术方案。例如上述特征与本公开中公开的(但不限于)具有类似功能的技术特征进行互相替换而形成的技术方案。
此外，虽然采用特定次序描绘了多种操作，但是这不应当理解为要求这些操作以所示出的特定次序或以顺序次序执行来执行。在一定环境下，多任务和并行处理可能是有利的。同样地，虽然在上面论述中包含了若干具体实现细节，但是这些不应当被解释为对本公开的范围的限制。在单独的实施例的上下文中描述的某些特征还可以组合地实现在单个实施例中。相反地，在单个实施例的上下文中描述的多种特征也可以单独地或以任何合适的子组合的方式实现在多个实施例中。

Claims (14)

  1. 一种图像拼接方法，包括：
    获取待拼接图像和多个模板图像，其中，所述待拼接图像和所述模板图像分别包括目标对象；
    计算所述待拼接图像与每个模板图像之间所述目标对象的匹配度；
    将所述匹配度满足预设匹配条件的模板图像作为目标模板图像；
    将所述待拼接图像中的目标图像部分拼接至所述目标模板图像中的目标图像区域内，得到目标拼接图像，所述目标图像部分和所述目标图像区域对应所述目标对象的相同部分。
  2. 根据权利要求1所述的方法，其中，所述计算所述待拼接图像与每个模板图像之间所述目标对象的匹配度，包括：
    确定所述待拼接图像中所述目标对象的待拼接对象特征点；
    针对所述每个模板图像，将所述模板图像中所述目标对象的模板对象特征点映射至所述待拼接图像上，得到所述模板图像对应的所述目标对象的映射对象特征点；
    根据所述待拼接对象特征点和映射对象特征点，计算所述待拼接图像与所述每个模板图像之间所述目标对象的匹配度。
  3. 根据权利要求2所述的方法，其中，所述将所述模板图像中所述目标对象的模板对象特征点映射至所述待拼接图像上，得到所述模板图像对应的所述目标对象的映射对象特征点，包括：
    计算所述模板对象特征点到所述待拼接对象特征点的最优仿射变换矩阵；
    根据所述最优仿射变换矩阵，将所述模板对象特征点分别映射至所述待拼接图像上，得到所述映射对象特征点。
  4. 根据权利要求2所述的方法，其中，所述匹配度包括所述映射对象特征点所属的像素和所述待拼接对象特征点所属像素的平均像素差；
    其中，所述匹配条件包括下列中的任一项：
    所述平均像素差在所述多个模板图像中最小；
    所述平均像素差小于或等于预设像素差阈值。
  5. 根据权利要求2所述的方法，其中，所述匹配度包括所述映射对象特征点所属的像素与所述待拼接对象特征点所属像素的平均距离；
    其中，所述匹配条件包括下列中的任一项：
    所述平均距离在所述多个模板图像中最小；
    所述平均距离小于或等于预设距离阈值。
  6. 根据权利要求1所述的方法，其中，所述将所述待拼接图像中的目标图像部分拼接至所述目标模板图像中的目标图像区域内，得到目标拼接图像，包括：
    将所述待拼接图像映射到所述目标模板图像上，得到映射后的目标模板图像；
    基于所述目标模板图像对应的拼接掩码图像，将所述映射后的目标模板图像中的目标图像部分拼接至所述目标模板图像中的目标图像区域内，得到所述目标拼接图像。
  7. 根据权利要求6所述的方法，其中，所述将所述待拼接图像映射到所述目标模板图像上，得到映射后的目标模板图像，包括：
    计算所述目标模板图像中所述目标对象的目标模板对象特征点到所述待拼接图像中所述目标对象的待拼接对象特征点的最优仿射变换逆矩阵；
    根据所述最优仿射变换逆矩阵，将所述待拼接图像映射到所述目标模板图像上，得到所述映射后的目标模板图像。
  8. 根据权利要求6所述的方法，在所述将所述待拼接图像中的目标图像部分拼接至所述目标模板图像中的目标图像区域内，得到目标拼接图像之前，所述方法还包括：
    根据所述目标模板对象特征点，确定所述目标图像区域；
    根据所述目标图像区域，生成所述拼接掩码图像。
  9. 根据权利要求1-8中任一项所述的方法，在所述获取待拼接图像和多个模板图像之前，所述方法还包括：
    获取多个预设图像；
    在所述多个预设图像中，筛选符合预设模板条件的所述多个模板图像；
    其中，所述预设模板条件包括下列中的至少一项：所述目标对象的图像占比大于或等于预设比值阈值；所述目标对象的对象角度小于或等于预设角度阈值；所述目标对象的对象尺寸大于或等于预设尺寸阈值。
  10. 根据权利要求1-8中任一项所述的方法，其中，所述目标对象包括脸部。
  11. 根据权利要求1所述的方法，其中，所述目标拼接图像的数量为多个；
    其中，在所述将所述待拼接图像中的目标图像部分拼接至所述目标模板图像中的目标图像区域内，得到目标拼接图像之后，所述方法还包括：
    根据预设视频模板和多个所述目标拼接图像，生成目标视频。
  12. 一种图像拼接装置，包括：
    第一获取单元，配置为获取待拼接图像和多个模板图像，其中，所述待拼接图像和所述模板图像分别包括目标对象；
    第一处理单元，配置为计算所述待拼接图像与每个所述模板图像之间的所述目标对象的匹配度；
    第二处理单元，配置为将所述匹配度满足预设匹配条件的模板图像作为目标模板图像；
    图像拼接单元，配置为将所述待拼接图像中的目标图像部分拼接至所述目标模板图像中的目标图像区域内，得到目标拼接图像，所述目标图像部分和所述目标图像区域对应所述目标对象的相同部分。
  13. 一种图像拼接设备，包括：
    处理器；
    存储器，用于存储可执行指令；
    其中，所述处理器用于从所述存储器中读取所述可执行指令，并执行所述可执行指令以实现上述权利要求1-11中任一项所述的图像拼接方法。
  14. 一种计算机可读存储介质，所述存储介质存储有计算机程序，当所述计算机程序被处理器执行时，使得处理器实现上述权利要求1-11中任一项所述的图像拼接方法。
PCT/CN2022/086739 2021-05-07 2022-04-14 图像拼接方法、装置、设备及介质 Ceased WO2022233223A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US18/558,048 US12555197B2 (en) 2021-05-07 2022-04-14 Splicing target object from input images to target template images

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202110496619.6A CN115311178A (zh) 2021-05-07 2021-05-07 图像拼接方法、装置、设备及介质
CN202110496619.6 2021-05-07

Publications (1)

Publication Number Publication Date
WO2022233223A1 true WO2022233223A1 (zh) 2022-11-10

Family

ID=83853631

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/086739 Ceased WO2022233223A1 (zh) 2021-05-07 2022-04-14 图像拼接方法、装置、设备及介质

Country Status (3)

Country Link
US (1) US12555197B2 (zh)
CN (1) CN115311178A (zh)
WO (1) WO2022233223A1 (zh)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20220358735A1 (en) * 2021-09-29 2022-11-10 Beijing Baidu Netcom Science Technology Co., Ltd. Method for processing image, device and storage medium
CN117173071A (zh) * 2023-11-02 2023-12-05 青岛天仁微纳科技有限责任公司 一种纳米压印模具的图像拼接方法
CN117541764A (zh) * 2024-01-09 2024-02-09 北京大学 一种图像拼接方法、电子设备及存储介质

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030007687A1 (en) * 2001-07-05 2003-01-09 Jasc Software, Inc. Correction of "red-eye" effects in images
CN104463081A (zh) * 2013-09-16 2015-03-25 展讯通信(天津)有限公司 人眼状态的检测方法
CN105118082A (zh) * 2015-07-30 2015-12-02 科大讯飞股份有限公司 个性化视频生成方法及系统
CN109559362A (zh) * 2018-11-23 2019-04-02 广东智媒云图科技股份有限公司 一种图像主体脸部替换方法及装置

Family Cites Families (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
GB0607143D0 (en) * 2006-04-08 2006-05-17 Univ Manchester Method of locating features of an object
JP5158223B2 (ja) * 2011-04-06 2013-03-06 カシオ計算機株式会社 三次元モデリング装置、三次元モデリング方法、ならびに、プログラム
CN107767330B (zh) * 2017-10-17 2021-02-26 中电科新型智慧城市研究院有限公司 一种图像拼接方法
CN110889314B (zh) * 2018-09-10 2022-09-13 北京市商汤科技开发有限公司 图像处理方法、装置、电子设备、服务器及系统
CN111008927B (zh) * 2019-08-07 2023-10-31 深圳华侨城文化旅游科技集团有限公司 一种人脸替换方法、存储介质及终端设备
KR102641424B1 (ko) * 2019-08-09 2024-02-28 삼성전자주식회사 컴퓨팅 장치 및 그 동작 방법
CN110458121B (zh) * 2019-08-15 2023-03-14 京东方科技集团股份有限公司 一种人脸图像生成的方法及装置
CN110852221B (zh) 2019-10-30 2023-08-18 深圳智慧林网络科技有限公司 基于块结合的人脸智能识别方法、终端和存储介质
CN111476710B (zh) * 2020-04-13 2022-12-02 上海艾麒信息科技有限公司 基于移动平台的视频换脸方法及系统
CN111861954B (zh) 2020-06-22 2024-10-18 北京百度网讯科技有限公司 编辑人脸的方法、装置、电子设备和可读存储介质
CN112508777A (zh) * 2020-12-18 2021-03-16 咪咕文化科技有限公司 一种美颜方法、电子设备及存储介质
US11544061B2 (en) * 2020-12-22 2023-01-03 International Business Machines Corporation Analog hardware matrix computation

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030007687A1 (en) * 2001-07-05 2003-01-09 Jasc Software, Inc. Correction of "red-eye" effects in images
CN104463081A (zh) * 2013-09-16 2015-03-25 展讯通信(天津)有限公司 人眼状态的检测方法
CN105118082A (zh) * 2015-07-30 2015-12-02 科大讯飞股份有限公司 个性化视频生成方法及系统
CN109559362A (zh) * 2018-11-23 2019-04-02 广东智媒云图科技股份有限公司 一种图像主体脸部替换方法及装置

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20220358735A1 (en) * 2021-09-29 2022-11-10 Beijing Baidu Netcom Science Technology Co., Ltd. Method for processing image, device and storage medium
CN117173071A (zh) * 2023-11-02 2023-12-05 青岛天仁微纳科技有限责任公司 一种纳米压印模具的图像拼接方法
CN117173071B (zh) * 2023-11-02 2024-01-30 青岛天仁微纳科技有限责任公司 一种纳米压印模具的图像拼接方法
CN117541764A (zh) * 2024-01-09 2024-02-09 北京大学 一种图像拼接方法、电子设备及存储介质
CN117541764B (zh) * 2024-01-09 2024-04-05 北京大学 一种图像拼接方法、电子设备及存储介质

Also Published As

Publication number Publication date
US12555197B2 (en) 2026-02-17
US20240221126A1 (en) 2024-07-04
CN115311178A (zh) 2022-11-08

Similar Documents

Publication Publication Date Title
CN114419300B (zh) 风格化图像生成方法、装置、电子设备及存储介质
WO2022166872A1 (zh) 一种特效展示方法、装置、设备及介质
CN109670444B (zh) 姿态检测模型的生成、姿态检测方法、装置、设备及介质
WO2022068487A1 (zh) 风格图像生成方法、模型训练方法、装置、设备和介质
CN111414879B (zh) 人脸遮挡程度识别方法、装置、电子设备及可读存储介质
CN111696176B (zh) 图像处理方法、装置、电子设备及计算机可读介质
WO2022233223A1 (zh) 图像拼接方法、装置、设备及介质
WO2023125374A1 (zh) 图像处理方法、装置、电子设备及存储介质
CN111369427A (zh) 图像处理方法、装置、可读介质和电子设备
WO2023093897A1 (zh) 图像处理方法、装置、电子设备及存储介质
CN110084154B (zh) 渲染图像的方法、装置、电子设备和计算机可读存储介质
CN111402122A (zh) 图像的贴图处理方法、装置、可读介质和电子设备
CN113610034B (zh) 识别视频中人物实体的方法、装置、存储介质及电子设备
US12526460B2 (en) Video generation method and apparatus, and device and storage medium
WO2023072015A1 (zh) 人物风格形象图的生成方法、装置、设备及存储介质
CN114422698A (zh) 视频生成方法、装置、设备及存储介质
CN114841984B (zh) 图像处理方法、装置、电子设备及存储介质
WO2020253716A1 (zh) 图像生成方法和装置
WO2024037556A1 (zh) 图像处理方法、装置、设备及存储介质
WO2023029991A1 (zh) 拍摄方法、装置、电子设备和存储介质
CN110619602B (zh) 一种图像生成方法、装置、电子设备及存储介质
CN112101258A (zh) 图像处理方法、装置、电子设备和计算机可读介质
CN116360661A (zh) 特效处理方法、装置、电子设备及存储介质
EP4170588B1 (en) Video photographing method and apparatus, and device and storage medium
CN111784726A (zh) 人像抠图方法和装置

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22798564

Country of ref document: EP

Kind code of ref document: A1

WWE Wipo information: entry into national phase

Ref document number: 18558048

Country of ref document: US

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 22798564

Country of ref document: EP

Kind code of ref document: A1

WWG Wipo information: grant in national office

Ref document number: 18558048

Country of ref document: US