Olympus Corporation

International Business Machines Corporation

Aya Nakashima

Osamu Koshiba

Hiromi Kobayashi

Masaharu Sakamoto

Yuki Satake

Yasuko Sonoda

Tatsuki Sawada

Atsushi Kohashi

Junichi Hanzawa

Kazuya Hirayu

Sho Ikawa

Mingqin Xu

A computer implemented method trains an image recognition model. A set of processor units creates a saliency map of an original image. The set of processor units superimposes the saliency map on the original image to form an augmented image, wherein the augmented image is used to train the image recognition model.

Data augmentation for machine learning

Automatically generating and implementing access policies is provided. An output of a needed user access context to access a resource is obtained from a trained machine learning model based on an extracted attribute of the resource input into the trained machine learning model. The access to the resource by a user is controlled using the output of the needed user access context to access the resource as an access policy for the resource.

Automatic resource access policy generation and implementation

A correlation value calculation device for selecting a pixel of interest within an image and a plurality of reference pixels within the image or a different image and calculating correlation values between a base block including the pixel of interest at a prescribed position and a plurality of reference blocks including each reference pixel at the prescribed position and having the same block size as the base block includes a block size determination unit, first intermediate correlation value calculation units configured to calculate correlation values in a plurality of types of block sizes, second intermediate correlation value calculation units configured to calculate correlation values in some block sizes among the plurality of types of block sizes, and a correlation value interpolation unit configured to obtain correlation values according to an interpolation.

Correlation value calculation device

A correlation value calculation device for calculating a correlation value between a base block within an image and at least one reference block within the image or other image includes a pixel value distribution calculation unit configured to calculate a pixel value distribution of pixels within the image and select a range of pixel values of a large number of pixels as a valid pixel value range, a conversion coefficient calculation unit configured to calculate conversion coefficients for converting pixels of the valid pixel value range into a prescribed bit-depth, and a pixel value conversion unit configured to perform a bit-depth reduction process on pixels within the base block and pixels within the reference block on the basis of the conversion coefficients.

A feature amount generation device includes a feature point detecting unit which detects a feature point from an input image, a pixel information generating unit which generates pixel information from pixel values of the surroundings of the feature point in the input image, a feature point pixel information distribution generating unit which generates a first distribution of pixel information of a first region including the feature point, a sub-feature point detecting unit which detects sub-feature points that are feature points in the surroundings of the feature point, a sub-feature point pixel information distribution generating unit which generates second distributions of pixel information of second regions including the sub-feature points, and a feature amount generating unit which generates a feature amount at the feature point based on the first distribution and the second distributions.

Growing community of inventors

Tokyo, Japan

Aya Nakashima