Perceptually-guided deep neural networks for ego-action prediction: Object grasping
作者：
Highlights：
• Biologically inspired models for grasping action prediction.
• Gaze-driven detection model for objects to be grasped in egocentric video.
• Two alternative methods for noise handling in eye-gaze measurements.
• A novel loss for automatic prediction of grasping actions.
• A new public dataset for prediction of grasping actions in ego-centric video.
摘要
•Biologically inspired models for grasping action prediction.•Gaze-driven detection model for objects to be grasped in egocentric video.•Two alternative methods for noise handling in eye-gaze measurements.•A novel loss for automatic prediction of grasping actions.•A new public dataset for prediction of grasping actions in ego-centric video.
论文关键词：Human perception,Grasping action prediction,Weakly supervised active object detection
论文评审过程：Received 31 March 2018, Revised 16 October 2018, Accepted 17 November 2018, Available online 17 November 2018, Version of Record 24 November 2018.
论文官网地址：https://doi.org/10.1016/j.patcog.2018.11.013
