Matterport, Inc.

Amazon Technologies, Inc.

Sony Corporation

Sony Electronics Inc

Gunnar Hovden

Mykhaylo Kurinnyy

Matthew Tschudy Bell

Volodymyr V Ivanchenko

Scott Adams

Kenneth Mark Karakotsios

David Alan Gausebeck

Isaac Scott Noble

Dong Zhou

Daniel Ford

Kah Kuen Fu

Steven Ka Cheung Moy

David V Buchhofer, Jr

Azwad Sabik

Techniques are provided for increasing the accuracy of automated classifications produced by a machine learning engine. Specifically, the classification produced by a machine learning engine for one photo-realistic image is adjusted based on the classifications produced by the machine learning engine for other photo-realistic images that correspond to the same portion of a 3D model that has been generated based on the photo-realistic images. Techniques are also provided for using the classifications of the photo-realistic images that were used to create a 3D model to automatically classify portions of the 3D model. The classifications assigned to the various portions of the 3D model in this manner may also be used as a factor for automatically segmenting the 3D model.

Automated classification based on photo-realistic image/model mappings

Systems, computer-implemented methods, apparatus and/or computer program products are provided that facilitate improving the accuracy of global positioning system (GPS) coordinates of indoor photos. The disclosed subject matter further provides systems, computer-implemented methods, apparatus and/or computer program products that facilitate generating exterior photos of structures based on GPS coordinates of indoor photos.

Selecting exterior images of a structure based on capture positions of indoor images associated with the structure

A system comprising: processors and memory containing instructions to control processors to: receive images representing an interior of a physical environment, identify, using neural network for object recognition, an object in an image, the object is associated with a location relative to the physical environment, identify, using neural network for object recognition, another object in another image, determine if objects in the images are located near or at a similar location based on location information associated with the objects, if the objects are located near or at a similar location, then objects are an instance of a single object, store similar location associated with the single object, display an interactive walkthrough visualization of a 3D model of the physical environment including the single object, receive request regarding object location through the interactive walkthrough visualization, and provide the similar location of the single object for display in the interactive walkthrough visualization.

System and method of object detection and interactive 3D models

Systems and techniques for determining and/or generating data for an architectural opening area associated with a three-dimensional (3D) model are presented. A portion of an image associated with a 3D model that corresponds to a window view or another architectural opening area is identified based at least in part on color data or depth data. Furthermore, a surface associated with the 3D model and visual data for the window view or the other architectural opening area is determined. The visual data for the window view or the other architectural opening area is applied to the surface associated with the 3D model.

Determining and/or generating data for an architectural opening area associated with a captured three-dimensional model

Techniques are disclosed for automated selection of navigation points for navigating through a computer-generated virtual environment such as, for example, a virtual reality (VR) environment. Specifically, an input set of connected navigation points in a virtual environment is automatically pruned to a connected subset thereof, according to one or more selection factors where at least one of the selection factors is whether the subset will continue to be 'connected' after pruning a navigation point from the input set. In addition to whether the pruning of a navigation point will allow the remaining navigation points to remain connected, techniques for pruning based on one or more additional selection factors are also disclosed. According to one additional selection factor, navigation point pruning is based on the degree to which coverage of an input set of navigation points would be reduced by pruning any given navigation point from the input set.

Navigation point selection for navigating through virtual environments

The use of resources on a computing device can be optimized for current conditions to reduce or minimize the amount of resources needed to provide a sufficient level of performance for various types of tasks. In some embodiments, one or more optimization algorithms can analyze information such as a type of task to be performed and the state of one or more environmental conditions to attempt to select a number and configuration of components, such as cameras and illumination elements, to use in performing the tasks. The performance of the tasks can be monitored, and the selection or states updated in order to maintain a sufficient level of performance without using more resources than are necessary.

Managing resource usage for task performance

A camera of a computing device can capture two or more images of a region including an object of interest, in order to allow for separation of the object from a background of the images through a process such as image subtraction. In order to compensate for rotations of the device between image captures, an element such as an electronic gyroscope can be used to monitor changes in orientation and predict an amount of shift of objects between images. The predicted shift can be used to attempt to align images captured around the time of the rotation, in order to enable subtraction or similar processes by effectively removing the shifting effect of the rotation.

Movement prediction for image capture

An electronic device automatically reverses a left-right orientation of content displayed by the device when detecting that a user is viewing content of the device via a mirror. Some embodiments reverse the content upon detecting a trigger such as a reflection of an image displayed by the device, indicating that the device is displaying content against a mirror. In some embodiments, the device reverses the content upon detecting that blinking markers displayed by the device are inverted or upon detecting a blinking marker flashing in a particular pattern consistent to that emitted by the device. Some embodiments reverse the content upon detecting both the reflection of the image and a reflection of the user. Further, some embodiments reverse the content upon determining that the user is facing a mirror and switches back to its normal orientation when the user is viewing the content by directly looking at the device.

Mirror detection-based device functionality

A system and method are disclosed for performing an image search procedure in which a histogram calculator initially performs image analysis procedures to construct horizontal and vertical texture histograms corresponding to texture characteristics of subject images. A feature detector may then perform the image search procedure by comparing the texture histograms to identify one or more matching images in which texture characteristics are most similar.

Growing community of inventors

Los Gatos, CA, United States of America

Gunnar Hovden