Hewlett-packard Development Company, L.p.

Hewlett-packard Company

Purdue Research Foundation

Other

Micro Focus Lcc

Qian Lin

Daniel R Tretter

Jan Philip Allebach

Jian Fan

Ullas Gargi

Rafael Dal Zotto

Clayton Atkins

Brian E Hoffmann

Suk Hwan Lim

Izhak Baharav

Ping Wah Wong

Pere Obrador

Jonathan Yen

Madhu Sudan Athreya

Feng Tang

Shaoyuan Xu

Malcolm Stuart Rix

Adam C Petruszka

Xiaoyu Xiang

Meichun Hsu

D Amnon Silverstein

Tong Zhang

Doron Shaked

Qiming Chen

Ramin Samadani

Charles Addison Bouman

Niranjan Damera-Venkata

Kent E Biggs

Alexander Wayne Clark

Jerry J Liu

Peng Wu

Harold Merkel

Yining Deng

Jeffrey L Trask

Georgia Koutrika

Andrew J Patti

Hugh Phu Nguyen

Guoxing Yang

Brent Michael Bradburn

Mei Chen

Konstantinos Konstantinides

Hao Tang

Kong Kritayakirana

David Murphy

Alexander Perumal, Jr

Nur Arad

Michael Stokes

Zhen He

Caio Northfleet

Andre Gobbi Farina

Thomas D Kite

Joao Luis Prauchner

Yang Cheng

Dun Liu

Gabriel Lima Paz

Guilherme Augusto Silva Megeto

Deangeli Gomes Neves

Augusto Cavalcante Valente

Weijuan Xi

Gabriel Lando

Madhu Sud Athreya

Doran Shaked

Neerja Raman

Hojohn Lee

Thiago Jose Da Silva

A saliency map of an image is generated. Saliency regions of the saliency map are identified. The saliency regions are merged into a combined saliency region. Candidate image crops of the image are generated based on the combined saliency region. An image crop of the image is selected from the candidate image crops using a machine learning model.

Image crop

An example system includes a feature extraction engine to identify features in a compressed image at a plurality of scales. The features include features corresponding to compression artifacts and features corresponding to image content to be upsampled. The feature extraction engine is to identify the features in the compressed image at a first scale based on noncontiguous pixels. The system also includes a reconstruction engine to refine the features corresponding to the image content to be upsampled and mitigate the features corresponding to the compression artifacts. The system includes an upsampling engine to generate an upsampled version of the compressed image based on the refined and mitigated features.

Image upsampling

A device includes a display panel and a processor. The processor is to receive information corresponding to an image displayed on the display panel and determine whether a relevant portion of the image is centered with respect to an area of interest of the display panel. The processor is further to responsive to a determination that the relevant portion of the image is not centered with respect to the area of interest of the display panel, perform an operation to cause the relevant portion of the image to be adjusted such that the relevant portion of the image is centered with respect to the area of interest of the display panel.

Image adjustment

An example device is described for facilitating polygon localization. In various aspects, the device can comprise a processor. In various instances, the device can comprise a non-transitory machine-readable memory that can store machine-readable instructions. In various cases, the processor can execute the machine-readable instructions, which can cause the processor to localize a polygon depicted in an image, based on execution of a deep learning pipeline. In various aspects, the deep learning pipeline can comprise a circular-softmax block.

Polygon localization via a circular-softmax block

An example system includes a first camera to capture a first image of a first facial body part and a second camera to capture a second image of a second facial body part. The example system further includes a transformation engine to transform respective scales of the first image and the second image to a scale of full facial images. The example system further includes a local location engine to: identify first facial landmarks and second facial landmarks in respective transformed versions of the first image and the second image, the first facial landmarks and the second facial landmarks used to determine that an action has occurred.

Two image facial action detection

An example system includes: a landmark detection engine to detect landmark positions of landmarks in images based on facial detection; an optical flow landmark engine to determine the landmark positions in the images based on optical flow of the landmarks between the images; a landmark difference engine to determine, for a landmark in a given image: a distance between a detected landmark position and an optical flow landmark position of the landmark; and a weighted landmark determination engine to determine, for a first and second image, a position for the landmark in the second image based on: a respective detected landmark position and a respective optical flow position of the landmark in the second image; and respective distances, determined with the landmark difference engine, between a first detected landmark position of the landmark in the first image and respective optical flow landmark positions for the first and second images.

Landmark temporal smoothing

Examples of methods for training using rendered images are described herein. In some examples, a method may include, for a set of iterations, randomly positioning a three-dimensional (3D) object model in a virtual space with random textures. In some examples, the method may include, for the set of iterations, rendering a two-dimensional (2D) image of the 3D object model in the virtual space and a corresponding annotation image. In some examples, the method may include training a machine learning model using the rendered 2D images and corresponding annotation images.

Training using rendered images

An example system includes a landmark engine to detect a facial landmark in an image of a face. The system includes a comparison engine to determine a difference between the facial landmark in the image and a facial landmark of a neutral face. The system also includes an action engine to determine whether a facial action unit occurred based on whether the difference satisfies a condition.

Facial action unit detection

In an example implementation according to aspects of the present disclosure, a device, method, and storage medium. The device comprises a processor and memory with instructions to receive a digital image and extract administrative and descriptive metadata. The device stores the administrative and descriptive metadata in a knowledge base and determines a digital image quality score. The device creates a recommendation based on the administrative and descriptive metadata and the digital image quality score.

Knowledge base image recommendations

In some examples an electronic device includes an image sensor and a processor. The processor is to determine an angle of an eye gaze of a user in relation to an optical axis of the image sensor and adjust, responsive to a determination that the angle exceeds a threshold angle, an active area of a display device.

Active areas of display devices

Examples disclosed herein provide the ability for a computing device to adjust settings on a computing device. In one example method, the computing device captures, via a first sensor of a computing device, images of an environment that the computing device is currently in, and detects objects in the environment as captured by the images. As an example, the computing device determines a location of the environment based on contextual data gathered from the detected objects and adjusts a setting on the computing device based on the determined location.

Adjusting settings on computing devices based on location

In example implementations, a method is provided. The method may be executed by a processor. The method includes receiving an image. The image is analyzed to obtain facial features. Contextual information is obtained and a vector including a facial feature class of the facial features and contextual feature classes of the contextual information is generated. A facial recognition is then performed based on the vector.

Facial recognitions based on contextual information

In one example in accordance with the present disclosure, a system is described. The system includes a pose estimator to identify a plurality of anatomical points on a person depicted in an image. A frame former of the system generates a frame for the person by connecting a first set of the plurality of anatomical points to form a skeleton and forming at least a head region and a torso region of the person based on a second set of the plurality of anatomical points. A map generator of the system generates a spatial gradient map projecting outwards from the frame. In the system, the spatial gradient map is based on pixel distance from the frame and an intensity along the gradient map indicates a pixels likelihood of form a part of the person.

Generating spatial gradient maps for a person in an image

In example implementations, a method is provided. The method may be executed by a processor. The method includes receiving an image. A person from a plurality of people within the image is identified. The person is segmented from the image. A background image of the image is replaced with a clean background image to remove the plurality of people from the image except the person that is identified.

Person segmentations for background replacements

Examples disclosed herein relate to a database comparison operation to identify an object. For example, a processor may enroll a set of object templates in a storage based on objects within input content and enroll a target object template in the storage based on a target object in target content. The processor may identify an object within the input content associated with the target object based on a database comparison operation of the stored set of object templates to the stored target object template. The processor may output object recognition information related to the identified object.

Database comparison operation to identify an object

Growing community of inventors

Santa Clara, CA, United States of America

Qian Lin