A method includes obtaining (such as accessing, receiving, acquiring, etc.), using at least one processor of an electronic device, a machine learning model trained to process input data and generate output data over at least one range of values associated with one or more control variables. The method also includes providing, using the at least one processor, specified input data to the machine learning model and providing, using the at least one processor, one or more specified values of the one or more control variables to the machine learning model. The one or more specified values of the one or more control variables are within the at least one range of values. The method further includes performing inferencing using the machine learning model to process the specified input data and generate specified output data. The inferencing is controlled based on the one or more specified values of the control variable(s).

One embodiment provides a method generating a first image crop and a second image crop randomly extracted from a low-quality image and a high-quality image, respectively. The method further comprises comparing the first image crop and the second image crop using a plurality of loss functions including pixel-wise loss to calculate losses, and optimizing a model trained to estimate a realistic scale-independent blur kernel of a low-resolution (LR) blurred image by minimizing the losses.

A computer-implemented method is provided for visualizing multiple objects in a computerized visual environment. The method includes displaying to a user a virtual three-dimensional space via a viewing device worn by the user, and determining a data limit of the viewing device for object rendering. The method includes presenting an initial rendering of the objects within the virtual space, where the visualization data used for the initial rendering does not exceed the data limit of the viewing device. The method also includes tracking user attention relative to the objects as the user navigates through the virtual space and determining, based on the tracking of user attention, one or more select objects from the multiple objects to which the user is paying attention. The one or more select objects are located within a viewing range of the user.

The present disclosure relates to systems, non-transitory computer-readable media, and methods that utilize a progressive refinement network to refine alpha mattes generated utilizing a mask-guided matting neural network. In particular, the disclosed systems can use the matting neural network to process a digital image and a coarse guidance mask to generate alpha mattes at discrete neural network layers. In turn, the disclosed systems can use the progressive refinement network to combine alpha mattes and refine areas of uncertainty. For example, the progressive refinement network can combine a core alpha matte corresponding to more certain core regions of a first alpha matte and a boundary alpha matte corresponding to uncertain boundary regions of a second, higher resolution alpha matte. Based on the combination of the core alpha matte and the boundary alpha matte, the disclosed systems can generate a final alpha matte for use in image matting processes.

A camera array, an imaging device and/or a method for capturing image that employ a plurality of imagers fabricated on a substrate is provided. Each imager includes a plurality of pixels. The plurality of imagers include a first imager having a first imaging characteristics and a second imager having a second imaging characteristics. The images generated by the plurality of imagers are processed to obtain an enhanced image compared to images captured by the imagers. Each imager may be associated with an optical element fabricated using a wafer level optics (WLO) technology.

The present application discloses a face super-resolution realization method and apparatus, an electronic device and a storage medium, and relate to fields of face image processing and deep learning. The specific implementation solution is as follows: a face part in a first image is extracted; the face part is input into a pre-trained face super-resolution model to obtain a super-sharp face image; a semantic segmentation image corresponding to the super-sharp face image is acquired; and the face part in the first image is replaced with the super-sharp face image, by utilizing the semantic segmentation image, to obtain a face super-resolution image.

Disclosed are a method of acquiring an image using a light-emitting element array and an apparatus therefor. The method of acquiring an image using a light-emitting element array includes reconstructing a first image from some images among source images, detecting a partial region containing a detection target object from the first image, acquiring partial-region images corresponding to the partial region from each of the source images, and reconstructing a second image from the partial-region images using the FPMP.

A method for microscope-based super-resolution includes acquiring a to-be-processed image and at least an auxiliary image, the to-be-processed image includes a target area, the auxiliary image includes an overlapping portion with the target area, and the to-be-processed image and the auxiliary image are both microscope images of a first resolution. The method further includes registering the to-be-processed image and the auxiliary image to obtain a registered image, and extracting one or more high-resolution features from the registered image. The one or more high-resolution features represent image features of the target area in a second resolution, and the second resolution is greater than the first resolution. The method also includes reconstructing, based on the one or more high-resolution features, a target image of the second resolution corresponding to the to-be-processed image of the first resolution. Apparatus and non-transitory computer-readable storage medium counterpart embodiments are also contemplated.

Systems and methods for multi-level vascular imaging for construction and display of vasculature from large to small vessels and micro-vessels using a combination of varying resolution contrast enhanced ultrasound flow imaging modalities are disclosed. While one or more resolution flow imaging modes may be employed for imaging large to small vessels of a vascular tree within a large region of interest, a high resolution mode, such as super resolution imaging, constructed for delineation of the microvascular morphology and directional microcirculation is provided within one or more small ROIs placed in selected locations within the larger ROI.

The present embodiment provides an image restoration method and apparatus which generate independent different restoration models by performing learning for each of different resolutions, receive a distorted image, and apply a restoration model corresponding to the resolution of the distorted image among the independent different restoration models to restore the distorted image into an improved upscaled image centering on a restoration target object within the distorted image.