Abstract: An image description information generation method includes obtaining a to-be-processed target image, and inputting the target image into a target-image description information generation network. The target-image description information generation network is a generation network that is obtained by performing adversarial training using a plurality of sample images and that is configured to generate image description information, the adversarial training is training an initialized image description information generation network and an initialized discriminative network alternately, and the discriminative network is configured to discriminate an output result of the image description information generation network. The method also includes according to the output result of the target-image description information generation network, generating target-image description information used for describing the target image.

Abstract: Embodiments of this disclosure disclose a face detection method, apparatus, and device, and a storage medium. A face candidate region in a to-be-detected image is determined according to a face detection model including a plurality of stages of convolutional neural networks (CNNs). A small-scale face is determined according to a size parameter, and a particular stage CNN corresponding to the size parameter is selected to perform the face detection. A CNN in different stage is adapted to a different size parameter or a size parameter range.

Abstract: An image description information generation method includes obtaining a to-be-processed target image, and inputting the target image into a target-image description information generation network. The target-image description information generation network is a generation network that is obtained by performing adversarial training using a plurality of sample images and that is configured to generate image description information, the adversarial training is training an initialized image description information generation network and an initialized discriminative network alternately, and the discriminative network is configured to discriminate an output result of the image description information generation network. The method also includes according to the output result of the target-image description information generation network, generating target-image description information used for describing the target image.