CN-122024020-A - Image-based human body measurement method, device and equipment

CN122024020ACN 122024020 ACN122024020 ACN 122024020ACN-122024020-A

Abstract

The embodiment of the application provides a human body measurement method, device and equipment based on images, and relates to the technical field of image processing. The method comprises the steps of obtaining image perception information of an image to be detected, determining a human body measurement task strategy corresponding to the image to be detected based on the image perception information, and executing human body measurement operation based on the image to be detected according to the human body measurement task strategy to obtain a human body measurement result. According to the embodiment of the application, the image sensing information is obtained, and the human body measurement task strategy is generated according to the image sensing information, so that the human body measurement of different types of images is realized in a targeted manner, and the universality, flexibility and robustness of the human body measurement are effectively improved.

Inventors

LIU CHEN
HE GAOFENG
WANG HUAMIN

Assignees

凌迪（浙江）科技股份有限公司

Dates

Publication Date: 20260512
Application Date: 20260415

Claims (11)

1. An image-based anthropometric method, comprising: acquiring image perception information of an image to be detected; determining a human body measurement task strategy corresponding to the image to be detected based on the image perception information, wherein the human body measurement task strategy comprises a plurality of measurement subtasks; performing human body measurement operation based on the image to be detected according to the human body measurement task strategy to obtain a human body measurement result; the image perception information comprises image quality evaluation information, and the acquisition mode of the image quality evaluation information comprises the following steps: And carrying out quality evaluation on the image to be detected from a plurality of preset evaluation index dimensions to obtain image quality evaluation information corresponding to the image to be detected.
2. The image-based anthropometric method of claim 1, wherein the image sensing information further comprises image semantic understanding information, and the image semantic understanding information is obtained by a method comprising the following steps: taking the image to be detected and the image quality evaluation information as inputs, and acquiring image semantic understanding information corresponding to the image to be detected by using a preset multi-mode visual language model; The image semantic understanding information comprises at least one of image semantic description information, human body identification influence factor information and initial task agenda information, wherein the image semantic description information is used for representing the basic condition of the image to be detected, the human body identification influence factor information is used for representing whether a preset degradation factor influencing human body identification exists in the image to be detected or not, and the initial task agenda information is used for representing the human body analysis task type suitable for the image to be detected which is preliminarily predicted.
3. The image-based anthropometric method of claim 1, wherein the determining a anthropometric task policy corresponding to the image to be detected based on the image perception information comprises: And comprehensively analyzing the image perception information and a preset task execution experience library to determine a human body measurement task strategy corresponding to the image to be detected, wherein the task execution experience library is used for storing record information of historical human body measurement tasks.
4. The image-based anthropometric method of claim 1, wherein the determining a anthropometric task policy corresponding to the image to be detected based on the image perception information comprises: And comprehensively analyzing the image sensing information and preset user configuration parameters to determine a human body measurement task strategy corresponding to the image to be detected, wherein the user configuration parameters are used for representing a measurement function starting state, measurement preference information and execution time constraint configured by a user.
5. The image-based anthropometric method of claim 1, wherein the performing anthropometric tasks based on the image to be detected according to the anthropometric task policies to obtain anthropometric results comprises: determining corresponding measurement subtasks and task execution sequences based on the anthropometric task policies; Respectively calling a task execution module corresponding to each measurement subtask according to the task execution sequence to execute the human body measurement operation, so as to obtain a measurement task result corresponding to each measurement subtask; and integrating based on the measurement task results to obtain a human body measurement result.
6. The image-based anthropometric method of claim 5, wherein the invoking the task execution modules corresponding to the measurement subtasks in the task execution order to execute the anthropometric tasks to obtain measurement task results corresponding to the measurement subtasks, respectively, comprises: Each task execution module executes human body measurement operation by adopting a preset mixed expert model to obtain a plurality of corresponding candidate task results; And respectively carrying out quality scoring on each candidate task result by using a preset scoring model, and taking the candidate task result with the highest quality score as a measurement task result corresponding to the task execution module.
7. The image-based anthropometric method of claim 6, further comprising: Judging that a task abnormal event occurs and generating corresponding abnormal description information under the condition that the quality score of a measurement task result corresponding to a target task execution module is lower than a preset score threshold value, wherein the target task execution module is any one of a plurality of task execution modules; Updating the anthropometric task strategy based on the anomaly description information; And adjusting and executing the human body measurement operation according to the updated human body measurement task strategy.
8. The image-based anthropometric method of claim 7, further comprising: And executing a preset fault-tolerant degradation strategy under the condition that the number of times of continuously generating the task abnormal events by the target task execution module is larger than a preset number of times threshold, wherein the fault-tolerant degradation strategy comprises reducing and adjusting the scoring threshold.
9. The image-based anthropometric method of any one of claims 1 to 8, wherein the step of obtaining image sensing information of an image to be detected is performed by a sensing agent, the step of determining a anthropometric task policy corresponding to the image to be detected based on the image sensing information is performed by a task scheduler agent, and the step of performing anthropometric tasks based on the image to be detected in accordance with the anthropometric task policy to obtain anthropometric results is performed by an executing agent.
10. An image-based anthropometric device, comprising: The information sensing module is used for acquiring image sensing information of the image to be detected; The strategy determining module is used for determining a human body measurement task strategy corresponding to the image to be detected based on the image perception information, wherein the human body measurement task strategy comprises a plurality of measurement subtasks; the measurement execution module is used for executing human body measurement operation based on the image to be detected according to the human body measurement task strategy to obtain a human body measurement result; the image perception information comprises image quality evaluation information, and the acquisition mode of the image quality evaluation information comprises the following steps: And carrying out quality evaluation on the image to be detected from a plurality of preset evaluation index dimensions to obtain image quality evaluation information corresponding to the image to be detected.
11. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the processor is operable to implement the image-based anthropometric method of any of claims 1-9 when the program is executed by the processor.

Description

Image-based human body measurement method, device and equipment Technical Field The present application relates to the field of image processing technologies, and in particular, to an image-based human body measurement method, apparatus, and device. Background In the fields of virtual fitting, body shape evaluation, health monitoring and the like, measurement of human body dimensions is one of the most basic tasks. Traditional tape measurement and measurement mode that utilizes professional equipment to carry out human scanning, though the detection precision is higher, can't satisfy the demand of measuring in different places under the network environment. The measurement mode of scanning or shooting images by using the mobile phone is more and more commonly applied because the mobile phone has the characteristics of low cost and high convenience and is suitable for remote measurement scenes in different places. However, existing image-based anthropometric schemes mainly detect based on fixed preset strategies or models, and lack of differential detection strategies for different types of images results in insufficient versatility, flexibility and robustness in anthropometric measurements based on images. Disclosure of Invention The embodiment of the application aims to provide an image-based human body measurement method, device and equipment, which are used for improving the universality, flexibility and robustness of human body measurement based on images. In a first aspect, an embodiment of the present application provides an image-based anthropometric method, including: acquiring image perception information of an image to be detected; determining a human body measurement task strategy corresponding to the image to be detected based on the image perception information, wherein the human body measurement task strategy comprises a plurality of measurement subtasks; and executing human body measurement operation based on the image to be detected according to the human body measurement task strategy to obtain a human body measurement result. In the embodiment of the application, the human body measurement task strategy is generated according to the image sensing information by acquiring the image sensing information, so that the human body measurement on different types of images is realized in a targeted manner, and the universality, flexibility and robustness of the human body measurement are effectively improved. In some embodiments, the image perception information comprises image quality evaluation information, and the image quality evaluation information is obtained by the following steps: And carrying out quality evaluation on the image to be detected from a plurality of preset evaluation index dimensions to obtain image quality evaluation information corresponding to the image to be detected. In the embodiment of the application, the accuracy of generating the measurement task strategy is further improved by acquiring the quality evaluation information of the image from multiple dimensions. In some embodiments, the image sensing information further comprises image semantic understanding information, and the acquiring mode of the image semantic understanding information comprises the following steps: taking the image to be detected and the image quality evaluation information as inputs, and acquiring image semantic understanding information corresponding to the image to be detected by using a preset multi-mode visual language model; The image semantic understanding information comprises at least one of image semantic description information, human body identification influence factor information and initial task agenda information, wherein the image semantic description information is used for representing the basic condition of the image to be detected, the human body identification influence factor information is used for representing whether a preset degradation factor influencing human body identification exists in the image to be detected or not, and the initial task agenda information is used for representing the human body analysis task type suitable for the image to be detected which is preliminarily predicted. In the embodiment of the application, the high-level semantic understanding information of the image is obtained by combining the quality evaluation information, so that the accuracy of generating the measurement task strategy is further improved. In some embodiments, the determining a anthropometric task policy corresponding to the image to be detected based on the image perception information comprises: And comprehensively analyzing the image perception information and a preset task execution experience library to determine a human body measurement task strategy corresponding to the image to be detected, wherein the task execution experience library is used for storing record information of historical human body measurement tasks. In the embodiment of the application, the human bod