Multi-target detection method, device and electronic device

Abstract

A method, apparatus and electronic device for multi-target detection are provide, wherein that multi-target detection method comprises: generate a plurality of detection frames according to the target; Acquiring classification probabilities of the plurality of detection frames; Generating confidence degrees of the plurality of detection frames according to classification probabilities of the plurality of detection frames, and obtaining stability degrees of the plurality of detection frames; And screening the plurality of detection frames according to confidence and stability of the plurality of detection frames. The method can effectively utilize the computing resources of the electronic device, detect meaningful targets in real time and stably through a detection frame, thereby reducing the resource occupancy of the server side, reducing the traffic consumption of users, and improving the user experience in related application scenarios.

Description

technical field [0001] The present application relates to the technical field of electronic equipment, in particular to a multi-target detection method, device and electronic equipment. Background technique [0002] At present, in order to ensure the accuracy of target detection, by deploying complex target detection models on the server side, when users use electronic devices to take pictures or videos, they can remotely access the detection service on the server side to obtain detection results. The detection results include Each object included in the picture or video frame is taken. [0003] However, in actual application, the applicant found that the above method of obtaining test results has the following technical problems: [0004] First, since electronic devices need to continuously upload images to the server side to obtain detection results, the acquisition of detection results greatly depends on the real-time nature of the network. In the case of network failure...

AI Technical Summary

Problems solved by technology

[0004] First, since electronic devices need to continuously upload images to the server side to obtain test results, the acquisition of test results greatly depends on the real-time performance of the network. There is obvious lag and consumes a lot of network traffic for users;

[0005] Second, increase the resource occupation on the server side. In scenarios such as users shooting videos, the server side needs to continuously detect video frames, which brings greater service request pressure on the server side.

Images