US 12,249,076 B2
Three-dimensional edge detection based on mutual object and edge detection and three-dimensional edge refinement detection
Luyan Liu, Shenzhen (CN); Kai Ma, Shenzhen (CN); and Yefeng Zheng, Shenzhen (CN)
Assigned to Tencent Technology (Shenzhen) Company Limited, Shenzhen (CN)
Filed by Tencent Technology (Shenzhen) Company Limited, Guangdong (CN)
Filed on Mar. 24, 2022, as Appl. No. 17/703,829.
Application 17/703,829 is a continuation of application No. PCT/CN2020/121120, filed on Oct. 15, 2020.
Claims priority of application No. 202010104850.1 (CN), filed on Feb. 20, 2020.
Prior Publication US 2022/0215558 A1, Jul. 7, 2022
Int. Cl. G06T 7/13 (2017.01); G06T 7/12 (2017.01); G06T 7/90 (2017.01); G06T 9/20 (2006.01); G06V 10/56 (2022.01); G06V 20/64 (2022.01)
CPC G06T 7/13 (2017.01) [G06T 7/12 (2017.01); G06T 7/90 (2017.01); G06T 9/20 (2013.01); G06V 10/56 (2022.01); G06V 20/647 (2022.01); G06T 2200/04 (2013.01); G06T 2207/10012 (2013.01); G06T 2207/10024 (2013.01); G06T 2207/20021 (2013.01); G06T 2207/20081 (2013.01)] 18 Claims
OG exemplary drawing
 
1. A method for three-dimensional edge detection, the method comprising: obtaining, for each of plural two-dimensional slices of a three-dimensional image, a two-dimensional object detection result and a two-dimensional edge detection result; stacking the two-dimensional object detection results into a three-dimensional object detection result, and stacking the two-dimensional edge detection results into a three-dimensional edge detection result E;
performing more than one encoding operation according to a feature map F of the three-dimensional image, the three-dimensional object detection result, and the three-dimensional edge detection result, to obtain an encoding result by
performing a first dot product operation ⊗ on the feature map of the three-dimensional image and the three-dimensional object detection result which is F⊗E, adding a result of the first dot product operation to the three-dimensional edge detection result, and performing a current encoding operation of the result of the first dot product operation added with the three-dimensional edge detection result; and
performing a second dot product operation on an output of the current encoding operation and the three-dimensional object detection result, adding a result of the second dot product operation to the three-dimensional edge detection result, and performing a next encoding operation of the second dot product operation added with the three-dimensional edge detection result, the more than one encoding operation that includes the current encoding operation and the next encoding operation being performed until a last encoding operation of the more than one encoding operation; and
performing, by processing circuitry of a computer device, decoding according to the encoding result, the three-dimensional object detection result, and the three-dimensional edge detection result, to obtain an optimized three-dimensional edge detection result of the three-dimensional image.