Method for detecting and tracking target object, target object tracking apparatus, and computer-program product
AssigneeBoe Technology Group Co., Ltd.
InventorYu Gu
Life of this patent
Estimating the term…
Abstract
The present application discloses a method performed by an electronic apparatus for detecting and tracking a target object. The method includes obtaining a first frame of scene; performing object detection and recognition of at least two portions of the target object respectively in at least two bounding boxes of the first frame of scene; obtaining a second frame of scene, the second frame of scene being later in time than the first frame of scene; and performing object tracking of the at least two portions of the target object respectively in the at least two bounding boxes of the first frame of scene.
Claims17
A method performed by an electronic apparatus for detecting and tracking a target object:
- wherein the electronic apparatus comprises an image sensor, a memory, and one or more processors, wherein the memory and the one or more processors are connected with each other;
- wherein the method comprises:
- obtaining a first frame of scene;
- performing object detection and recognition of at least two portions of the target object respectively in at least two bounding boxes of the first frame of scene, wherein the at least two portions of the target object comprises one or more feature-rich portions of the target object respectively in one or more feature-rich bounding boxes;
- obtaining a second frame of scene, the second frame of scene being later in time than the first frame of scene;
- performing object tracking of the at least two portions of the target object respectively in the at least two bounding boxes of the first frame of scene;
- extracting a plurality of feature points from the at least two portions of the target object in the first frame of scene, the one or more feature-rich bounding boxes having a higher distribution density of feature points than other bounding boxes;
- and;
- performing object tracking of the at least two portions of the target object respectively in the at least two bounding boxes based on the plurality of feature points in the first frame of scene.
A target object tracking apparatus, comprising:
- a memory and one or more processors, wherein the memory and the one or more processors are connected with each other, the memory stores computer-executable instructions for controlling the one or more processors to perform object detection and recognition of at least two portions of the target object respectively in at least two bounding boxes of a first frame of scene;
- at least two object trackers configured to perform object tracking of the at least two portions of the target object respectively in the at least two bounding boxes of the first frame of scene;
- and
- one or more image sensors configured to obtain a first frame of scene and a second frame of scene, the second frame of scene is later in time than the first frame of scene;
- wherein the at least two portions of the target object comprises one or more feature-rich portions of the target object respectively in one or more feature-rich bounding boxes;
- the target object tracking apparatus further comprises an extractor configured to extract a plurality of feature points from the at least two portions of the target object in the first frame of scene, the one or more feature-rich bounding boxes having a higher distribution density of feature points than other bounding boxes;
- and
- the at least two object trackers is configured to perform object tracking of the at least two portions of the target object respectively in the at least two bounding boxes based on the plurality of feature points in the first frame of scene.
A computer-program product, comprising a non-transitory tangible computer-readable medium having instructions thereon, the instructions comprising:
- code for causing a target object tracking apparatus to obtain a first frame of scene;
- code for causing a target object tracking apparatus to perform object detection and recognition of at least two portions of the target object respectively in at least two bounding boxes of the first frame of scene, wherein the at least two portions of the target object comprises one or more feature-rich portions of the target object respectively in one or more feature-rich bounding boxes;
- code for causing a target object tracking apparatus to obtain a second frame of scene, the second frame of scene is later in time than the first frame of scene;
- code for causing a target object tracking apparatus to perform object tracking of the at least two portions of the target object respectively in the at least two bounding boxes of the first frame of scene;
- code for extracting a plurality of feature points from the at least two portions of the target object in the first frame of scene, the one or more feature-rich bounding boxes having a higher distribution density of feature points than other bounding boxes;
- and
- code for performing object tracking of the at least two portions of the target object respectively in the at least two bounding boxes based on the plurality of feature points in the first frame of scene.
Dependencies are inferred from the claim text.