Kyocera Document Solutions Inc.

Sheng Li

Dongpei Su

Kenneth Allen Schmidt

Michael M Chang

Kendrick Esperanza Wong

Stephen L Schafer

Alfred Abkarian

Larry Lubman

Guo Li

Yuzuru Naganuma

Kilho Shin

Thien-Phuc Nguyen Do

Waqas Rasheed

A device such as a color printer includes a main memory, a cache memory, and a convolutional neural network configured to convert pixels from a first color space to a second color space. The convolutional neural network is organized into execution-separable layers, and loaded one or more layer at a time (depending on cache size) from the main memory to the cache memory, whereby the pixels are processed through each of the layers in the cache memory, and layers that have completed processing are evicted to make room for caching next layer(s) of the network.

Complex color space conversion using convolutional neural networks

A system for super-sampling digital images detects artifacts in an SRGAN super-sampled image, determines blocks of the image that contribute to the artifacts, and if the artifact-contributing blocks exceed a threshold, discards the SRGAN generated output image in favor of applying a super-sampled image generated by an alternate mechanism, such as a nearest neighbor algorithm.

Mitigation of quantization-induced image artifacts

A digital image processor includes a region proposal network configured to transform digital image inputs into region proposals and bounding box refinement logic configured to transform the region proposals by determining a first set of the region proposals exhibiting dense spacing, determining a second set of the region proposals exhibiting sparse spacing, executing a first transformation to merge at least some of the region proposals exhibiting dense spacing to generate refined region proposals, executing a second transformation to join at least some of the region proposals exhibiting sparse spacing to generate additional ones of the refined region proposals, and applying an expansion transformation to the refined region proposals.

Consolidation of bounding boxes generated from text type pattern localizer

Systems and methods upscale an input image by a final upscaling factor. The systems and methods employ a first module implementing a super resolution neural network with feature extraction layers and multiple sets of upscaling layers sharing the feature extraction layers. The multiple sets of upscaling layers upscale the input image according to different respective upscaling factors to produce respective first module outputs. The systems and methods select the first module output with the respective upscaling factor closest to the final upscaling factor. If the respective upscaling factor for the selected first module output is equal to the final upscaling factor, the systems and methods output the selected first module output. Otherwise, the systems and methods provide the selected first module output to a second module that upscales the selected first module output to produce a second module output corresponding to the input image upscaled by the final upscaling factor.

Super resolution neural network with multiple outputs with different upscaling factors

Systems and methods for processing images receive an input image. The systems and methods provide the input image to a first module to increase a resolution of the input image to produce an upscaled image. The systems and methods detect white pixels in the input image. The systems and methods generate a mask associated with the input image. The mask includes mask bits that are set to mark the white pixels in the input image. The systems and methods upscale the mask to produce an upscaled mask matching a resolution of the upscaled image. The systems and methods identify target pixels of the upscaled image that correspond to the set mask bits in the upscaled mask. The systems and methods modify the upscaled image to produce an output image by replacing target pixels of the upscaled image with a replacement pixel having greater whiteness. The systems and methods output the output image.

White background protection in SRGAN based super resolution

An example system includes a processor and a non-transitory computer-readable medium having stored therein instructions that are executable to cause the system to perform various functions. The functions include obtaining an image associated with a print job, and providing the image as input to a convolutional neural network. The convolutional neural network includes a residual network, upscaling layers, and classification layers configured to detect whether the image is an artificial image having a computer-generated image gradient. The functions also include determining, based on an output of the classification layers, that the image is an artificial image having a computer-generated image gradient. Further, the functions include, based on determining that the image is an artificial image having a computer-generated image gradient, providing the image to an upscaling module of a print pipeline for upscaling rather than using an output of the upscaling layers for the upscaling.

Super-resolution convolutional neural network with gradient image detection

Example systems and related methods relate to rendering using smooth shading. A computing device can receive information about an image, the information including one or more patches, where a particular patch of the one or more patches is specified using a plurality of non-linear equations. The computing device can determine one or more linear approximations to a particular non-linear equation of the plurality of non-linear equations. The computing device can update the particular patch to replace the particular non-linear equation with at least one linear approximation of the one or more linear approximations. The computing device can render at least part of the image by at least rendering the updated particular patch. The computing device can generate an output that includes the rendered at least part of the image.

Utilizing smooth shading patches in image rendering

An example embodiment may involve obtaining an a&#xd7;b pixel macro-cell from an input image. Pixels in the a&#xd7;b pixel macro-cell may have respective pixel values and may be associated with respective tags. It may be determined whether at least e of the respective tags indicate that their associated pixels represent edges in the input image. Based on this determination, either a first encoding or a second encoding of the a&#xd7;b pixel macro-cell may be selected. The first encoding may weigh pixels that represent edges in the input image heavier than pixels that do not represent edges in the input image, and the second encoding might not consider whether pixels represent edges. The selected encoding may be performed and written to a computer-readable output medium.

Cell-based compression with edge detection

An example embodiment may involve obtaining an m&#xd7;n pixel cell from an input image. Each of the m&#xd7;n pixels in the m&#xd7;n pixel cell may be associated with at least one color value. An m&#xd7;n attribute cell may be determined, elements of which may be associated in a one-to-one fashion with respective pixels in the m&#xd7;n pixel cell. The m&#xd7;n pixel cell may be compressed in a lossy fashion, and the m&#xd7;n attribute cell may be compressed in a lossless fashion. Compression of the m&#xd7;n pixel cell may be based on at least part of the m&#xd7;n attribute cell. An interleaved representation of the compressed m&#xd7;n pixel cell and the compressed m&#xd7;n attribute cell may be written to an output medium.

Digital image attribute plane compression

An example embodiment may involve obtaining (i) an a&#xd7;b attribute macro-cell, and (ii) a&#xd7;b pixel macro-cells for each of a luminance plane, a first color plane, and a second color plane of an input image. The a&#xd7;b pixel macro-cells may each contain 4 non-overlapping m&#xd7;n pixel cells. The example embodiment may also involve determining 4 attribute-plane output values that represent the 4 non-overlapping m&#xd7;n attribute cells, 1 to 4 luminance-plane output values that represent the a&#xd7;b pixel macro-cell of the luminance plane, a first color-plane output value to represent the a&#xd7;b pixel macro-cell of the first color plane, and a second color-plane output value to represent the a&#xd7;b pixel macro-cell of the second color plane. The example embodiment may further involve writing an interleaved representation of the output values to a computer-readable output medium.

Cell-based compression with edge detection and interleaved encoding

An example embodiment may involve obtaining an a&#xd7;b pixel macro-cell from an input image. The a&#xd7;b pixel macro-cell may contain 4 non-overlapping m&#xd7;n pixel cells. The a&#xd7;b pixels in the a&#xd7;b pixel macro-cell may have respective color values and may be associated with respective object type tags. The example embodiment may also include selecting a compression technique to either (i) compress the a&#xd7;b pixel macro-cell as a whole, or (ii) compress the a&#xd7;b pixel macro-cell by compressing each of the 4 non-overlapping m&#xd7;n pixel cells separately. The example embodiment may further include compressing the a&#xd7;b pixel macro-cell according to the selected compression technique, and writing a representation of the compressed a&#xd7;b pixel macro-cell to a computer-readable output medium.

Digital image color plane compression

An example embodiment may involve obtaining an a&#xd7;b pixel macro-cell from an image with one or more color planes, and an a&#xd7;b attribute macro-cell. The a&#xd7;b pixel macro-cell may contain 4 non-overlapping m&#xd7;n pixel cells, and the a&#xd7;b attribute macro-cell may contain 4 non-overlapping m&#xd7;n attribute cells. The pixels in the a&#xd7;b pixel macro-cell may be associated with respective color values. The example embodiment may also involve determining 4 attribute output values associated respectively with the 4 non-overlapping m&#xd7;n attribute cells. The example embodiment may further involve determining 1 to 4 color-plane output values for the non-overlapping m&#xd7;n pixel cells, and writing an interleaved representation of the 4 attribute output values and the determined color-plane output values.

Interleaved encoding of compressed attribute and color planes

A computer-implemented method comprises receiving a data stream that includes a series of code words that encodes a respective series of pixel data according to a first entropy coding lookup table, and processing the data stream to determine if there is a match between a first code word and a consecutive second code word, and a code word entry in a second entropy coding lookup table. The method also includes, if there is a match, decoding the first code word and the second code word using the second entropy coding lookup table. Further, the method includes, if there is not a match, decoding the first code word using the first entropy coding lookup table.

Growing community of inventors

El Segundo, CA, United States of America

Sheng Li