The disclosure relates to a method for the noise optimization of a camera, in particular a handheld thermal imaging camera. Images are captured by means of the camera in at least one method step; at least one movement characteristic variable is detected by means of at least one sensor unit of the camera in at least one method step; and image data of captured images is averaged by means of a computing unit of the camera in at least one method step. At least a number of images to be averaged are determined by means of the computing unit of the camera at least on the basis of an intensity of the detected movement characteristic variable, in particular a change rate, in at least one method step.

Temporal filtering operations may be reset for certain pixels within an image frame to reduce contribution from previous input frames to reduce ghosting and other artifacts. The resetting reduces the contribution to, for example, zero, either immediately or within a predetermined period of time (e.g., a certain number of frames). A decision regarding whether to reset temporal filtering for a pixel of the image frame may be based on a probability assigned to that pixel. The probability can be based on rules with one or more criteria. One example factor for adjusting probability is a confidence level regarding the temporal filtering decision for the pixel, in which the probability for a random reset of a pixel is based on the confidence level regarding the temporal filtering decision for those pixels.

The invention provides methods and systems for generating an ultrasound image. In a method, the generation of an ultrasound image comprises: obtaining channel data, the channel data defining a set of imaged points; for each imaged point: isolating the channel data; performing a spatial spectral estimation on the isolated channel data; and selectively attenuating the spatial spectral estimation channel data, thereby generating filtered channel data; and summing the filtered channel data, thereby forming a filtered ultrasound image. In some examples, the method comprises aperture extrapolation. The aperture extrapolation improves the lateral resolution of the ultrasound image. In other examples, the method comprises transmit extrapolation. The transmit extrapolation improves the contrast of the image. In addition, the transmit extrapolation improves the frame rate and reduces the motion artifacts in the ultrasound image. In further examples, the aperture and transmit extrapolations may be combined.

The present disclosure is directed to systems and methods of multiphase pseudo-continuous arterial spin labeling.

A modular architecture is provided for denoising Monte Carlo renderings using neural networks. The temporal approach extracts and combines feature representations from neighboring frames rather than building a temporal context using recurrent connections. A multiscale architecture includes separate single-frame or temporal denoising modules for individual scales, and one or more scale compositor neural networks configured to adaptively blend individual scales. An error-predicting module is configured to produce adaptive sampling maps for a renderer to achieve more uniform residual noise distribution. An asymmetric loss function may be used for training the neural networks, which can provide control over the variance-bias trade-off during denoising.

Provided herein may be an image sensing device and a method of operating the same. An image sensing device may include an image sensor obtaining an image using a plurality of pixels, and an image processor configured to use pixel values included in a region of interest included in the image to generate a gain table including gain table values corresponding to a first resolution, convert the gain table into a target table including target table values corresponding to a second resolution which is lower than the first resolution, and cancel noise included in the image based on the target table.

Systems and methods are disclosed to manipulate or normalize image of animals to a reference size and shape. Synthetically normalizing the image data to a reference size and shape allows machine learning models to automatically identify subject behaviors in a manner that is robust to changes in the size and shape of the subject. The systems and methods of the invention can be applied to drug or gene therapy classification, drug or gene therapy screening, disease study including early detection of the onset of a disease, toxicology research, side-effect study, learning and memory process study, anxiety study, and analysis in consumer behavior.

The image processing method according to the present application includes: acquiring a medical image captured by an imaging apparatus; and determining an intensity of a filter to be applied to the medical image according to a degree of focusing of the medical image.

Apparatuses, systems, and techniques to generate blue noise masks for real-time image rendering and enhacement. In at least one embodiment, a noise mask is generated and applied to one or more images to generate one or more enhanced images for image processing (e.g., real-time image rendering). In at least one embodiment, the noise mask is able to handle the temporal domain (e.g., add time to the spatial domain) to improve image quality when rendering images over multiple frames.

An image analyzer unit performs lesion site detection for each tomographic image. In response to a lesion site detection state, the image analyzer unit outputs a position data array and a size data array that represent temporal change in position and size of the lesion site. A rate of change calculator unit calculates a rate of change that represents a degree of temporal change in tomographic image content. A smoother unit smooths the position data array and the size data array in accordance with the rate of change. A mark is generated based on the smoothed position data array and the smoothed size data array.