公开(公告)号：WO2021081568A2

公开(公告)日：2021-04-29

申请号：PCT/US2021/019240

申请日：2021-02-23

Applicant: INNOPEAK TECHNOLOGY, INC.

Inventor： SUN, Hongyu ,  LI, Chen ,  LI, Chengeng ,  ZHANG, Zhebin ,  YE, Xiaoyu ,  THORNTON, Daniel

IPC: G06T15/00

Abstract: Systems and methods are provided for an advanced stereoscopic 3D rendering system to solve several of these 3D rendering issues and work independently of different game engines. For example, the system can include a stereoscopic render mechanism to apply post-processing effects to OpenGL ES applications. The advanced stereoscopic 3D rendering system can be added as an interception layer between the game engine and the display screen of the user. This interception layer can be integrated with many different game engines to create a 3D view for various user device models, thus removing the need for image generators at the user device to create 3D images. The 3D images can be created by the advanced stereoscopic 3D rendering system for viewing by the end user and seemingly incorporated with the game or other software application.

公开(公告)号：WO2022182445A1

公开(公告)日：2022-09-01

申请号：PCT/US2022/013261

申请日：2022-01-21

Applicant: INNOPEAK TECHNOLOGY, INC.

Inventor： LIN, Ruiyuan ,  SUN, Hongyu ,  ZHANG, Zhebin ,  LIAO, Rijun ,  CAI, Binghuang ,  SUN, Jian

IPC: G06V10/74 ,  G06V10/44 ,  G06V10/82

Abstract: Novel tools and techniques are provided for implementing duplicate image or video determination and/or image or video deduplication based on deep metric learning with keypoint features. In various embodiments, a computing system may perform feature discriminability enhancement, using a deep metric learning framework, by: transforming first and second image-specific aggregated feature sets into first and second feature vectors using corresponding neural networks; converting each of the first and second feature vectors into first and second normalized feature vectors having values between 0 and 1, using an activation function; applying a loss function to the first and second normalized feature vectors to decrease Euclidean distance between positive pairs and increase Euclidean distance between negative pairs; and binarizing the first and second normalized feature vectors; and determining whether the second image is a duplicate of the first image based on a Hamming distance between the first and second binarized normalized feature vectors.

公开(公告)号：WO2022104296A1

公开(公告)日：2022-05-19

申请号：PCT/US2021/065473

申请日：2021-12-29

Applicant: INNOPEAK TECHNOLOGY, INC.

Inventor： YE, Xiaoyu ,  ZHANG, Zhebin ,  SUN, Hongyu ,  SUN, Jian

IPC: G01S13/86 ,  B60R1/00 ,  G01S13/93 ,  G01S13/931

Abstract: Novel tools and techniques are provided for implementing camera radar fusion for advanced driver assistance system ("ADAS") with radar and mobile device. In various embodiments, a computing system on a mobile device may receive one or more first images from a first camera that is mounted to a first position on a windshield of a first vehicle; may receive first radar data from a first radar sensor that is mounted behind the windshield of the first vehicle; may fuse the received one or more first images and the received first radar data to generate first fused data; may analyze the generated first fused data to perform object detection and tracking to identify, highlight, and track one or more first objects that are located in front of the first vehicle; and may present, on a display device, the identified, highlighted, and tracked one or more first objects.

公开(公告)号：WO2022104294A1

公开(公告)日：2022-05-19

申请号：PCT/US2021/065467

申请日：2021-12-29

Applicant: INNOPEAK TECHNOLOGY, INC.

Inventor： DONG, Xihua ,  ZHANG, Zhebin ,  SUN, Hongyu ,  SUN, Jian

IPC: H04N7/00

Abstract: Novel tools and techniques are provided for implementing advanced driver assistance system ("ADAS") with a camera(s) on a windshield of a vehicle and a mobile device. In various embodiments, a computing system on a mobile device may receive one or more images from a camera that is mounted to a fixed position on a windshield of a vehicle, may analyze the received one or more images to identify and highlight one or more objects captured by the camera, may generate one or more fused images by fusing the identified and highlighted one or more objects with the one or more images, and may analyze the one or more fused images to identify one or more alert conditions associated with operation of the vehicle. Based on a determination that at least one first alert condition has been identified, the computing system may perform one or more driver assistance tasks.

公开(公告)号：WO2022099224A1

公开(公告)日：2022-05-12

申请号：PCT/US2021/065167

申请日：2021-12-24

Applicant: INNOPEAK TECHNOLOGY, INC.

Inventor： XIAO, Gaoyu ,  ZHANG, Zhebin ,  SUN, Hongyu ,  SUN, Jian

IPC: G01S13/931 ,  G01S7/02 ,  G01S13/00

Abstract: Novel tools and techniques are provided for implementing radar-based object detection (e.g., advanced driver assistance systems ("ADASs"), other radar-based object detection, etc.). In various embodiments, a computing system may receive Radar data from a Radar sensor, the Radar sensor being disposed such that a transparent or translucent material is positioned between one or more objects and the Radar sensor. The computing system may parse the Radar data to extract range or depth data, and if at least one extracted range or depth data corresponds to a distance between the Radar sensor and the transparent or translucent material, may filter out at least one Radar data point corresponding to said at least one extracted range or depth data from the Radar data to produce filtered Radar data, and may utilize the filtered Radar data as input to an object detection system.

公开(公告)号：WO2022250690A1

公开(公告)日：2022-12-01

申请号：PCT/US2021/034783

申请日：2021-05-28

Applicant: INNOPEAK TECHNOLOGY, INC.

Inventor： SUN, Hongyu ,  LI, Chen ,  LI, Chengeng ,  ZHANG, Zhebin ,  CHON, Kevin

IPC: G06K9/46 ,  G06N3/08 ,  G06N5/02 ,  G06T7/40

Abstract: This application is directed to rendering graphic content from associated content data. An electronic device executes a user application configured to render the graphic content on a display screen. The electronic device obtains content data to be converted to the graphic content in the user application. A plurality of semantic features is extracted from the content data and applied to determine a semantic category associated with the content data based on the plurality of semantic features. In accordance with the semantic category, the electronic device renders the graphic content from the content data for display on the display screen. In some embodiments, a plurality of graphic elements is identified from the content data and rendered in the graphic content according to the semantic category.

公开(公告)号：WO2022104295A1

公开(公告)日：2022-05-19

申请号：PCT/US2021/065468

申请日：2021-12-29

Applicant: INNOPEAK TECHNOLOGY, INC.

Inventor： WU, Hsiang-Huang ,  ZHANG, Zhebin ,  SUN, Hongyu ,  SUN, Jian

IPC: B60W50/14 ,  B60W40/04 ,  G08B21/10 ,  H04W4/44 ,  H04W4/38

Abstract: Novel tools and techniques are provided for implementing object detection using fusion of vision, location, and/or other signals, and/or implementing alert condition messaging based thereon. In various embodiments, based on a determination that an alert condition(s) is near a location of a vehicle(s) or its predicted path(s), an alert condition message (including information regarding the alert condition and its determined location) may be sent or pushed to a computing system or user device associated with (and in some cases, located within or on) the vehicle(s). As the vehicle(s) approaches the alert condition location(s), its computing system or user device may fuse the alert condition with one or more of object detection data signals, alert condition location, and/or current location and/or current predicted path of the vehicle to generate fused data, which may then be used to generate and present alert messages indicating that the vehicle(s) is(are) approaching the alert condition.

公开(公告)号：WO2022099225A1

公开(公告)日：2022-05-12

申请号：PCT/US2021/065541

申请日：2021-12-29

Applicant: INNOPEAK TECHNOLOGY, INC.

Inventor： CHON, Kevin ,  ZHANG, Zhebin ,  SUN, Hongyu ,  SUN, Jian

IPC: H04N7/00

Abstract: The present invention is directed to point cloud generation. In a specific embodiment, the present invention provides a method for generating a cloud. Two-dimensional images captured by cameras are processed to generate initial point clouds. Using data received from a radar, an integrated point cloud is generated using the initial point clouds and the radar data. There are other embodiments as well.