Example-based affine registration is provided. In various embodiments, a plurality of training images is read. A predetermined affine transform is read for each of the plurality of training images. Each affine transform maps its associated image to a template. Weights are determined for each of the plurality of training images. The weights are determined to minimize a difference between the test image and a weighted linear combination of the training images. An affine transform is determined mapping the test image to the template by computing a weighted linear combination of the affine transforms using the weights.

An image processing apparatus cuts out a plurality of background images and foreground images from a plurality of images, and stores a plurality of cutout foreground images and an alignment coefficient calculated from an image prior to cutout of each foreground image, the plurality of foreground images being stored in association with the alignment coefficient. Moreover, the image processing apparatus generates a background combined image by combining cutout background images, and selects any of the stored foreground images based on designation of a position in the background combined image. Then, image processing apparatus determines an alignment coefficient of the selected foreground image by using the alignment coefficient associated with the selected foreground image, and combines the selected foreground image and the background combined image.

An image pickup apparatus includes an optical system and at least one processor executing instructions to: control the optical system to pick up images while changing a focus position; combine; and detect feature points of the images picked up by the optical system and use the feature points of two images to calculate a conversion coefficient for positioning. In the controlling, in a case where the focus positions of the two images are adjacent to each other and the conversion coefficient of the two images does not satisfy a predetermined condition, image pickup is performed again with a method for controlling change of the optical system, and in the combining, two images in which the conversion coefficient satisfies the predetermined condition and focus positions of which are adjacent to each other among the images picked up by the optical system are combined.

The invention relates to a registration apparatus (14) for registering images comprising a unit (11) for providing a first and a second image of an object, such that an image element of the first image at a respective position has been reconstructed by multiplying projection data values of rays traversing the image element with weights and by backprojecting the weighted projection data values, a unit (12) for providing a confidence map comprising for different positions in the first image confidence values being indicative of a likelihood that an image feature is caused by a structure of the object, the confidence value being calculated as a sum of a function, which depends on the respective weight, over the rays traversing the respective image element, and a unit (13) for determining a transformation for registering the first and second image to each other under consideration of the confidence map.

The present disclosure provides a method and device for stitching wind turbine blade images, and a storage medium. The method includes performing edge detection on a plurality of images of the blade of the wind turbine to determine a blade region for each of the plurality of images; and for each pair of images among the plurality of images of the blade of the wind turbine, which are captured successively, stitching a front end of a former one of the pair of images captured successively and a rear end of a latter one of the pair of images captured successively, wherein the front end is far away from a root of the blade of the wind turbine, and the rear end is close to the root of the blade of the wind turbine.

Systems and techniques for propagating spot healing edits from a source image to a target image include receiving a source image depicting a face with a healing region and a target image depicting the face. The face is detected in the source image including detecting facial feature points in the source image. The face is identified in the target image including detecting facial feature points in the target image. Facial feature point correspondence is determined between the facial feature points of the source image in the facial feature points of the target image. Region correspondence is determined between regions of the source image and regions of the target image using the facial feature point correspondence. The healing region of the face in the source image is transformed to a corresponding region of the face in the target image using the region correspondence.

A meme generation method, an electronic device, and a storage medium are provided. The method includes: determining a plurality of second expression images corresponding to a target face image based on a plurality of first expression images contained in a first meme; generating a second meme corresponding to the target face image based on the plurality of second expression images corresponding to the target face image; wherein, determining an affine transformation parameter between the target face image and an i-th first expression image in the plurality of first expression images according to a corresponding relation between a face key point in the target face image and a face key point in the i-th first expression image; and transforming the target face image based on the affine transformation parameter to obtain an i-th second expression image corresponding to the target face image.

Disclosed are a panoramic video forward mapping method and a panoramic video inverse mapping method, which relates to the field of virtual reality (VR) videos. In the present disclosure, the forward mapping method comprises: mapping, based on a main viewpoint, the Areas I, II, and III on the sphere onto corresponding areas on the plane, wherein Area I corresponds to the area with the included angle 0Z.sub.1, the Area II corresponds to the area with the included angle Z.sub.1Z.sub.2, and the Area III corresponds to the area with the included angle Z.sub.2180. The panoramic video forward mapping method refers to mapping a spherical source corresponding to the panoramic image A onto a plane square image B; the panoramic video inverse mapping method refers to mapping the plane square image B back to the sphere for being rendered and viewed. the present disclosure may significantly lower the resolution of a video, effectively lower the code rate for coding the panoramic video and reducing the complexity of coding and decoding, further achieving the objective of lowering the code rate and guaranteeing video quality of the ROI area.

Embodiments can provide a method for atlas-based contouring, comprising constructing a relevant atlas database; selecting one or more optimal atlases from the relevant atlas database; propagating one or more atlases; fusing the one or more atlases; and assessing the quality of one or more propagated contours.

Example-based affine registration is provided. In various embodiments, a plurality of training images is read. A predetermined affine transform is read for each of the plurality of training images. Each affine transform maps its associated image to a template. Weights are determined for each of the plurality of training images. The weights are determined to minimize a difference between the test image and a weighted linear combination of the training images. An affine transform is determined mapping the test image to the template by computing a weighted linear combination of the affine transforms using the weights.