Guangdong Oppo Mobile Telecommunications Corp., Ltd.

Innopeak Technology, Inc.

Yi Xu

Shuxue Quan

Yang Zhou

Yuliang Guo

Zhong Li

Yuxin Ma

Xiangyu Du

Yuan Tian

Yu Gao

Jun Liu

A method for identifying human poses in an image, a computer system, and a non-transitory computer-readable medium are provided. The method includes obtaining the image, encoding the image into a feature map, converting the feature map into a plurality of TPDF maps corresponding to a plurality of predefined part types, and generating a global pose map based on the feature map and the plurality of TPDF maps corresponding to the predefined part types. Each TPDF map corresponds to a respective predefined part type and represents a plurality of vectors connecting each position on the image to a closest body part among the one or more body parts of the respective predefined part type. A global pose map is generated based on the feature map and the plurality of TPDF maps, identifying one or more human bodies having a plurality of body parts in a scene of the image.

Method for identifying human poses in an image, computer system, and non-transitory computer-readable medium

The present disclosure provides a target detection method. The method includes: acquiring a scene image of a scene; acquiring a three-dimensional point cloud corresponding to the scene; identifying a supporting plane in the scene based on the three-dimensional point cloud; generating a plurality of region proposals based on the scene image and the supporting plane; and performing a target detection on the plurality of region proposals to determine a target object to be detected in the scene image. In addition, the present disclosure also provides a target detection device, a terminal device, and a medium.

Target detection method, device, terminal device, and medium

The present disclosure provides an object detection method. The method includes: acquiring a scene image of a scene; acquiring a three-dimensional point cloud corresponding to the scene; segmenting the scene image according to the three-dimensional point cloud corresponding to the scene to generate a plurality of region proposals; and performing object detection on the plurality of region proposals to determine a target object to be detected in the scene image. In addition, The present disclosure also provides an object detection device, a terminal device, and a medium.

Object detection method, object detection device, terminal device, and medium

An object detection method is provided. The method includes: acquiring a scene image of a scene; acquiring a three-dimensional point cloud corresponding to the scene; segmenting the scene image into a plurality of sub-regions; merging the plurality of sub-regions according to the three-dimensional point cloud to generate a plurality of region proposals; and performing object detection on the plurality of region proposals to determine a target object to be detected in the scene image. In addition, an object detection device, a terminal device, and a medium are also provided.

This application is directed to identifying user gestures. One or more first images are captured by a first camera and include a first perspective of a hand. A first hand gesture is identified from the one or more first images. A second hand gesture is identified from one or more second images that are captured by a second camera, e.g., substantially concurrently with the one or more first images. The one or more second images include a second perspective of the hand that is distinct from the first perspective. A final hand gesture of the hand is determined based on the first and second hand gestures and applied to enable an operation on an object. The first and second cameras are included in two distinct electronic devices. The operation is performed by one of the two electronic devices or by a third electronic device distinct from the two electronic devices.

Multiple perspective hand tracking

Methods and systems are disclosed for mapping the coordinate systems of devices for AR applications. A third device may receive a first pose from a first device associated with a first coordinate system and a second pose from a second device associated with a second coordinate system. The third device may receive an image that includes fiducial markers presented by each of the first device and the second device. The first pose and the portion of the image including the first fiducial marker may define a first set of 3D coordinates. The second pose and the portion of the image including the second fiducial marker may define a second set of 3D coordinates. A coordinate-system transform can be generated from a correspondence between the first set and the second of 3D coordinates. The coordinate-system transform may be transmitted to the first device and/or the second device.

Associating device coordinate systems in a multi-person augmented reality system

The present disclosure provides a computer-implemented method. The method includes: acquiring 3D volumetric data for representing a real-world scene; acquiring a current set of active 3D volumetric data from the 3D volumetric data in response to an activating action on the 3D volumetric data; updating a plurality of sets of active 3D volumetric data based on the current set of active 3D volumetric data, the plurality of sets of active 3D volumetric data being acquired from the 3D volumetric data; constructing a surface based on the plurality of updated sets of active 3D volumetric data; and simulating a deformation of a part of the real-world scene based on the surface, the part of the real-world scene corresponding to the plurality of updated sets of active 3D volumetric data. In addition, the present disclosure also provides an electronic device, and a storage medium.

Computer-implemented methods for simulating deformation in real-world scene, electronic devices, and computer-readable storage medium

An exemplary method of detecting a light source using an electronic device having a camera and a sensor may include: scanning a real environment using the camera to establish an environment map of the real environment; capturing, using the camera, a first image of a real light source from a first location in the real environment and a second image of the real light source from a second location in the real environment; tracking, using the sensor, a first position and a first orientation of the camera in the environment map while the first image is captured, and a second position and a second orientation of the camera in the environment map while the second image is captured; and computing a position of the real light source in the environment map based on the first position, the first orientation, the second position, and the second orientation.

Light source detection for extended reality technologies

A method for human body model reconstruction and a reconstruction system are disclosed. The method includes acquiring a target image, and acquiring a segmented image by segmenting the target image based on an object to be reconstructed in the target image, the target image being one front image of the object; acquiring an initial estimate shape and a part of texture information of the object respectively, according to the segmented image; determining an initial 3D model of the object through the initial estimate shape, the initial 3D model being a 3D model without texture; acquiring complete texture information of the object according to the part of texture information and a texture generation model; and generating a 3D reconstruction model of the object based on the initial 3D model and the complete texture information, the 3D reconstruction model being a 3D model with texture.

Method for human body model reconstruction and reconstruction system

A method includes: sensing, by a mobile device when the mobile device is tethered to a head-mounted display (HMD) and when the mobile device is held by a first hand, first data; detecting, by the mobile device, information indicating which hand is used to perform a gesture from the first data; sending, by the mobile device, the information to the HMD; receiving, by the HMD, the information, and not detecting, by the HMD, the information by an image-based detection method; wherein the gesture is performed by a second hand and the information is an updated indication with respect to the gesture considering a hand switch state during a first elapsed duration from a sensing time of the first data to a sensing time of the first image data; and performing, by the HMD, side-adapted hand pose estimation on the first image data using the information.

Method, mobile device, head-mounted display, and system for estimating hand pose

Techniques for rendering a 3D virtual object in an augmented-reality system are described. A system, a method, and a non-transitory memory device for augmented reality rendering of three-dimensional, virtual objects are described. In an example, a number of images of an environment are acquired; relative movement of a camera acquiring the number of images is tracked; camera pose is determined relative to the environment using the number of images and tracked relative movement of the camera; depth and normal surfaces of objects in the environment are estimated using a depth map and a normal map; a surface geometry of the environment is reconstructed using the depth map and the normal map; and the virtual object is rendered using the surface geometry of the environment.

Growing community of inventors

Palo Alto, CA, United States of America

Yi Xu