Dotproduct LLC

Rafael Spring

Mark Klusza

Thomas Greaves

A method for controlling optional constraints to processing of multi-dimensional scene data via a user interface [UI] in an image management device is disclosed. The first step in this process is receiving a first data set of a scene having location information about a first location in an image wherein the first data set has a first performance metric. Next is activating a Constraint Manager having a plurality of constraint processes. The next step is selecting a first Constraint process from the plurality of constraint processes. Then receiving a new data set for the first constraint process to apply to the first data set, before finally activating the first Constraint process to incorporate the new data set to estimate a new location data set for the first location, wherein the new location data set has an improved performance metric as compared to the first performance metric.

Semi-automatic 3D scene optimization with user-provided constraints

Augmentation of captured 3D scenes with contextual information is disclosed. A 3D capture device is used to capture a plurality of 3D images at a first resolution. A component on a mobile computing device is used to capture at least one piece of contextual information that includes a capture location data and a pose data. The mobile computing device receives, the plurality of 3D images from the 3D capture device, and renders the plurality of 3D images into a 3D model. In addition, the at least one piece of contextual information is embedded into a correct location in the 3D model. A user interactive version of the 3D model including the embedded at least one piece of contextual information is then displayed.

Augmentation of captured 3D scenes with contextual information

A system and method for real-time or near-real time processing and post-processing of RGB-D image data using a handheld portable device including extracting gray values from the RGB-D image data, creating image pyramids from the grey values and the depth data, computing a scene fitness value using the image pyramids, predicting a camera pose and aligning the image with a first subset of selected key frames, computing a new camera pose estimate and creating a keyframe using the new camera pose estimate, after the keyframe is created, selecting a second subset of keyframes different from the first subset of keyframes and repeating the step of aligning with each keyframe of the second subset of keyframes, deciding whether new links are required to the keyframe in a keyframe pose graph and linking the keyframes.

Handheld portable optical scanner and method of using

A system and methods for real-time or near-real time processing and post-processing of RGB-D image data using a handheld portable device and using the results for a variety of applications. The disclosure is based on the combination of off-the-shelf equipment (e.g. an RGB-D camera and a smartphone/tablet computer) in a self-contained unit capable of performing complex spatial reasoning tasks using highly optimized computer vision algorithms. New applications are disclosed using the instantaneous results obtained and the wireless connectivity of the host device for remote collaboration. One method includes steps of projecting a dot pattern from a light source onto a plurality of points on a scene, measuring distances to the points, and digitally reconstructing an image or images of the scene, such as a 3D view of the scene. A plurality of images may also be stitched together to re-position an orientation of the view of the scene.

Growing community of inventors

Wiesbaden, Germany

Rafael Spring