Target detection method, terminal equipment and storage medium

The invention discloses a target detection method. The method comprises the following steps: acquiring image data acquired by an image acquisition device; taking the features extracted in each stage of the residual network as the input of a preset recursive feature pyramid model; and obtaining feedb...

Full description

Saved in:
Bibliographic Details
Main Authors LIU HONG, SONG PINHAO, DING RUNWEI, DAI LINHUI
Format Patent
LanguageChinese
English
Published 24.08.2021
Subjects
Online AccessGet full text

Cover

More Information
Summary:The invention discloses a target detection method. The method comprises the following steps: acquiring image data acquired by an image acquisition device; taking the features extracted in each stage of the residual network as the input of a preset recursive feature pyramid model; and obtaining feedback information when the preset recursive feature pyramid model takes the features as input, and correcting the residual network according to the feedback information. The invention further discloses terminal equipment and a computer readable storage medium, and the effect of improving the robustness of the target recognition model is achieved. 本发明公开了一种目标检测方法,包括以下步骤:获取图像采集装置采集的图像数据;将残差网络各个阶段提取的特征作为预设的递归特征金字塔模型的输入;获取所述预设的递归特征金字塔模型以所述特征为输入时的反馈信息,并根据所述反馈信息修正所述残差网络。本发明还公开了一种终端设备及计算机可读存储介质,达成了提高目标识别模型的鲁棒性的效果。
Bibliography:Application Number: CN202110550672