Preference prediction based on a photo gallery analysis with scene recognition and object detection

作者：

Highlights：

• User preferences can be efficiently predicted from photo gallery using scene recognition and object detection.

• Representation of image by using scores and embeddings from the scene recognition model combined with the scores of object detector improves the accuracy of visual event recognition.

• Representations of all photos are combined into a single descriptor of a user with the attention mechanism in order to increase performance of set-based image recognition.

摘要

•User preferences can be efficiently predicted from photo gallery using scene recognition and object detection.•Representation of image by using scores and embeddings from the scene recognition model combined with the scores of object detector improves the accuracy of visual event recognition.•Representations of all photos are combined into a single descriptor of a user with the attention mechanism in order to increase performance of set-based image recognition.

论文关键词：Scene recognition,Event recognition,Object detection,Recognition of a set of images,CNN (Convolutional neural network),Mobile device

论文评审过程：Received 28 October 2019, Revised 5 August 2021, Accepted 11 August 2021, Available online 13 August 2021, Version of Record 19 August 2021.

论文官网地址：https://doi.org/10.1016/j.patcog.2021.108248