Image double-pyramid multivariate feature extraction network, image segmentation method, system and medium

The invention provides an image double-pyramid multivariate feature extraction network, an image segmentation method, a system and a medium and belongs to the field of deep learning image processing, which are composed of four input features, an instance feature pyramid, a semantic feature pyramid and two output features. The two output features are composed of instance feature pyramid output features and semantic feature pyramid output features. According to the method, the problem that a traditional feature extraction method cannot meet the feature requirement of a multi-thread task is solved, detailed instance target feature information can be provided for a target recognition task, rich semantic logic feature information can be provided for a semantic analysis task, and the precision of the multi-thread task is greatly improved. 图像的双金字塔多元特征提取网络、图像分割方法、系统和介质，属于深度学习图像处理领域，由四个输入特征、一个实例特征金字塔、一个语义特征金字塔和两个输出特征构成，两个输出特征由实例特征金字塔输出特征和语义特征金字塔输出特征组成。本发明解决了传统的特征提取方法无法满足多线程任务的特征需求的问题，能够为目标识别任务，提供详细的实例目标特征信息、为语义分析任务，提供