Systems for image evaluation and dynamic cropping are provided. In some examples, a system, may receive an instrument or image of an instrument. Identifying information may be extracted from the instrument or image of the instrument. Based on the extracted identifying information, a check/check image profile may be retrieved. In some examples, expected size and/or shape data may be extracted from the check/check image profile. The extracted expected size and/or shape data may be compared to size and/or shape data from the received instrument or image of the instrument to identify any anomalies (e.g., to determine whether the expected size and/or shape data matches the size and/or shape data of the received instrument or image of the instrument. If the expected size and/or shape data does not match size and/or shape data from the received instrument or image of the instrument, the instrument or image of the instrument may be programmatically modified and a modified image of the instrument may be generated.

Apparatus and methods relate to receiving an input image comprising an array of pixels, wherein the input image is associated with a first characteristic; applying a neural network to transform the input image to an output image associated with a second characteristic by generating, by an encoder and for each pixel of the array of pixels of the input image, an encoded pixel, providing, to a decoder, the array of encoded pixels, applying, by the decoder, axial attention to decode a given pixel, wherein the axial attention comprises a row attention or a column attention applied to one or more previously decoded pixels in rows or columns preceding a row or column associated with the given pixel, wherein the row or column attention mixes information within a respective row or column, and maintains independence between respective different rows or different columns; and generating, by the neural network, the output image.

A system for image completion is disclosed. The system comprises a coordinate generation module configured to receive past frames and a present frame having a first field-of-view and to generate a set of coordinate maps, one for each of the received past frames; and a frame aggregation module configured to receive as input the past frames, the present frame, and the coordinate maps and to synthesize, based on said input, a present frame having a second field-of-view.

An image processing method configured to correct a captured image using a neural network includes a first step of determining an inversion axis for a partial image or a filter of the neural network according to a position in the captured image of the partial image that is part of the captured image, a second step of determining a positional relationship among pixels of color components in an image input to the neural network corresponding to the inversion axis, and a third step of generating a corrected image obtained by correcting an input image by processing, using the neural network, the input image generated from the captured image based on the positional relationship.

The present invention provides an efficient image classification method based on structured pruning, which incorporates a spatial pruning method based on variation regularization, including steps such as image data preprocessing, inputting images to neural network, image model pruning and retraining, and new image class predication and classification. The present invention adopts a structured pruning method that removes unimportant weight parameters of the original network model and reduces unnecessary computational and memory consumptions caused by the network model in image classification to simplify the image classifier, and then uses the sparsified network model to predict and classify new images. The simplified method according to the present invention improves the original network model in image classification efficiency by nearly two times, costs about 30% less memory consumption and produces a better classification result.

A computational method is disclosed for producing a sequence of high-resolution (HR) images from an input sequence of low-resolution (LR) images. The method uses a structured subspace framework to learn pairs of LR dictionaries from the input LR sequence ‘and’ employ learned pairs of LR dictionaries into estimating HR images. The structured subspace framework itself is based on a pair of specially structured HR basis matrices, wherein a HR basis spans any HR image whose so-called polyphase components (PPCs) are spanned by the corresponding LR dictionary.

An image processing apparatus includes a determination unit configured to determine a region of the image on which to perform character recognition processing, a decision unit configured to decide, based on a number of black pixels in contact with the region determined by the determination unit, whether to perform the character recognition processing on an expanded region obtained by expanding the region determined by the determination unit rather than on the region determined by the determination unit, and a character recognition unit configured to perform the character recognition processing on that region of the image decided by the decision unit.

A hardware downscaler and an architecture for implementing a FIR filter in which the downscaler can be arranged for downscaling by a half in one dimension. The downscaler can comprise: hardware logic implementing a first three-tap FIR filter; and hardware logic implementing a second three-tap FIR filter; wherein the output from the hardware logic implementing the first three-tap filter is provided as an input to the hardware logic implementing the second three-tap filter.

A image processing method includes a shifting step for determining whether to shift an identical color pixel region, in which two to N pixels of an identical color are arranged consecutively in a first direction, in the first direction or a direction opposite to the first direction within the image data, provided that N is an integer of 2 or greater, and shifting the identical color pixel region in the first direction or the direction opposite to the first direction when the shifting is determined to be performed, and an averaging step, in which in the shifting step, the shifting is determined to be performed when the identical color pixel region exists across two of the unit regions, and the identical color pixel region spanning across the two of the unit regions is shifted so as to be included in any one of the unit regions.

A hardware downscaling module and downscaling methods for downscaling a two-dimensional array of values. The hardware downscaling unit comprises a first group of one-dimensional downscalers; and a second group of one-dimensional downscalers; wherein the first group of one-dimensional downscalers is arranged to receive a two-dimensional array of values and to perform downscaling in series in a first dimension; and wherein the second group of one-dimensional downscalers is arranged to receive an output from the first group of one-dimensional downscalers and to perform downscaling in series in a second dimension.