An eyewear camera as a system is introduced that includes many subsystems. These subsystems include: scene imaging, control methods, tracking a user's eye, methods and techniques to increase the resolution of an image captured by a scene camera, methods to create a viewfinder, and methods to capture an image of a user's eye while simultaneously projecting an image into the user's eye. Such an eyewear will allow a user to capture a scene effortlessly, select an object within the scene, get extra information about the object via a digital personal assistant, or even modify a subset of the scene. Specific applications will be discussed that include visual aid for people with low vision and upgrading existing security cameras via proposed single camera super resolution techniques.

A method of generating an output image having an output resolution of N pixels×N pixels, each pixel in the output image having a respective color value for each of a plurality of color channels, the method comprising: obtaining a low-resolution version of the output image; and upscaling the low-resolution version of the output image to generate the output image having the output resolution by repeatedly performing the following operations: obtaining a current version of the output image having a current K×K resolution; and processing the current version of the output image using a set of convolutional neural networks that are specific to the current resolution to generate an updated version of the output image having a 2K×2K resolution.

The present disclosure provides a method of training a generative adversarial network. The method includes iteratively enhancing a first noise input in a generative network to generate a first output image; iteratively enhancing a second noise input in the generative network to generate a second output image; transmitting the first output image and a second reference image to a discriminative network, the second reference image corresponding to the first reference image and having a higher resolution than the first reference image; obtaining a first score from the discriminative network based on the second reference image, and a second score from the discriminative network based on the first output image; calculating a loss function of the generative network based on the first score and the second score; and adjusting at least one parameter of the generative network to lower the loss function of the generative network.

A method of joint denoising and demosaicing using a neural network to generate an output image and to a computing device for implementing the method is provided. The method includes obtaining image data collected by a color filter array (CFA), and jointly performing denoising and demosaicing on the CFA image data using a trained neural network to generate an output image, wherein the neural network has a lightweight U-Net architecture and is trained on a plurality of pairs of training images. One image in each pair of training images is obtained with a lower ISO value than another image in the pair of training images, and is processed by a processing algorithm (image signal processor (ISP)), and the other image in each pair of training images is in a format of CFA image data.

A method comprising training a pre-trained artificial neural network using degraded data together with higher-quality reference data to obtain an adapted artificial neural network.

An approach for providing a cap with an embedded high-resolution lens and a sampling insert that is used during an authentication of a composition of a liquid in a container sealed by the cap. The cap has a top portion of a cap with an opening, a sampling insert inside the opening in the top portion of the cap, and a high-resolution lens inside an opening in the sampling insert.

A method of generating an output image having an output resolution of N pixels×N pixels, each pixel in the output image having a respective color value for each of a plurality of color channels, the method comprising: obtaining a low-resolution version of the output image; and upscaling the low-resolution version of the output image to generate the output image having the output resolution by repeatedly performing the following operations: obtaining a current version of the output image having a current K×K resolution; and processing the current version of the output image using a set of convolutional neural networks that are specific to the current resolution to generate an updated version of the output image having a 2K×2K resolution.

Techniques related to generating a fine super resolution image from a low resolution image including a person wearing a predetermined uniform are discussed. Such techniques include applying a pretrained convolutional neural network to a stacked image including image channels from a coarse super resolution image, label data corresponding to the coarse super resolution image from available labels relevant to the uniform, and pose data corresponding to the person to determine the fine super resolution image.

Embodiments of the present disclosure provide a solution for image/video super resolution. A method for image processing is proposed. The method comprises: receiving a first image with a first resolution and at least one reference image associated with the first image, the first image and the at least one reference image being associated with a same video; determining a difference between the first image and the at least one reference image; and generating a second image with a second resolution based on the difference, the first image and the at least one reference image, the second resolution being higher than the first resolution.

Methods and systems for performing a hybrid machine learning method for enhancing scanning electron microscopy (SEM) images are disclosed herein. Methods include the steps of acquiring a plurality of images of a region of a sample that were each generated by irradiating the sample with a pulsed charged particle beam, upscaling each of the individual images to generate a plurality of upscaled images of the region of the sample, and combining the plurality of upscaled images to form a noise reduced image of the region of the sample.