(EN) The invention discloses an image processing method, and the method comprises the steps: obtaining a face image which comprises an original face image and a to-be-recognized face image; merging the face images; transmitting the preset number of face images after merging processing from a memory corresponding to a central processing unit CPU to a video memory corresponding to a graphics processing unit GPU; inputting the preset number of face images after merging processing into a deep learning framework for calculation to obtain feature vectors corresponding to the preset number of face images;merging the feature vectors corresponding to the preset number of face images; and returning the feature vectors corresponding to the preset number of face images after merging processing from a video memory corresponding to a graphics processing unit GPU to a memory corresponding to a central processing unit CPU. The invention further provides an image processing device, a terminal and a computer readable storage medium. According to the method, the mode of moving the face images in batches is adopted, the waiting time of the GPU can be shortened, and the computing power of the GPU is improved.
(ZH) 一种图像处理方法,包括:获取人脸图像,人脸图像包括原始人脸图像与待识别的人脸图像;将人脸图像进行合并处理;将合并处理后的预设数量的人脸图像由中央处理器CPU对应的内存传递至图形处理器GPU对应的显存中;将合并处理后的预设数量的人脸图像输入至深度学习框架中计算,得到预设数量的人脸图像对应的特征向量;将预设数量的人脸图像对应的特征向量进行合并处理;将合并处理后的预设数量的人脸图像对应的特征向量由图形处理器GPU对应的显存回传至中央处理器CPU对应的内存中。本发明还提供一种图像处理装置、终端与计算机可读存储介质。本发明采用批量搬移人脸图像的方式,能够减少GPU的等待时间,提高GPU计算能力。