Intel Corporation

Varjo Technologies Oy

Joni-Matti Määttä

Jarno Nikkanen

Ville Timonen

Petteri Timonen

Ari Antti Erik Peuhkurinen

Noam Levy

German Voronov

Dmitry Rudoy

Edmond Chalom

Oleksandr Dovzhenko

Roman Golovanov

Tuomas Tölli

Harish Essaky Sankaran

Lauri Suomela

Tarek Mohsen

A method including: receiving visible-light images captured using camera(s) and depth data corresponding to said images; identifying image segments of visible-light image that represent objects or their parts belonging to different material categories; detecting whether at least two adjacent image segments in visible-light image pertain to at least two different material categories related to same object category; and when it is detected that at least two adjacent image segments pertain to at least two different material categories related to same object category, identifying at least two adjacent depth segments of depth data corresponding to at least two adjacent image segments; and correcting errors in optical depths represented in at least one of at least two adjacent depth segments, based on optical depths represented in remaining of at least two adjacent depth segments.

Depth maps and 3D reconstruction with segmentation masks

An imaging system including visible-light camera(s), depth sensor(s), pose-tracking means, and server(s) configured to: control visible-light camera(s) and depth sensor(s) to capture visible-light images and depth images of real-world environment, respectively, whilst processing pose-tracking data to determine poses of visible-light camera(s) and depth sensor(s); reconstruct three-dimensional lighting model of real-world environment representative of lighting in different regions of real-world environment; receive, from client application, request message comprising information indicative of location in real-world environment where virtual object(s) is to be placed; utilise three-dimensional lighting model to create sample lighting data for said location, wherein sample lighting data is representative of lighting at given location in real-world environment; and provide client application with sample lighting data.

Imaging systems and methods for facilitating local lighting

A system, article, and method to perform deep learning-based automatic white balancing uses a neural network in conjunction with other non-neural network automatic white balancing algorithms.

Method and system of deep learning-based automatic white balancing

Methods, apparatuses and systems may provide for operating a machine learning device by obtaining training image data, conducting an offline prediction analysis of the training image data with respect to one or more real-time parameters of an image capture device, and generating one or more parameter detection models based on the offline prediction analysis. Additionally, methods, apparatuses and systems may provide for operating the image capture device by obtaining a candidate image associated with the image capture device, determining that the candidate image corresponds to a particular type of scene represented in a parameter prediction model, and adjusting one or more real-time parameters of the image capture device based at least in part on one or more parameter values associated with the particular type of scene.

Machine learning of real-time image capture parameters

Growing community of inventors

Tampere, Finland

Joni-Matti Määttä