SaltiNet: Scan-path Prediction on 360 Degree Images using Saliency Volumes

Abstract

We introduce SaltiNet, a deep neural network for scanpath prediction trained on 360-degree images. The model is based on a temporal-aware novel representation of saliency information named the saliency volume. The first part of the network consists of a model trained to generate saliency volumes, whose parameters are fit by back-propagation computed from a binary cross entropy (BCE) loss over downsampled versions of the saliency volumes. Sampling strategies over these volumes are used to generate scanpaths over the 360-degree images. Our experiments show the advantages of using saliency volumes, and how they can be used for related tasks. Our source code and trained models available at https://github.com/massens/saliency-360salient-2017.

Winner of the Best Scan-path Award at the Salient360!: Visual attention modeling for 360 degrees Images Grand Challenge of ICME 2017. Presented at the ICCV 2017 Workshop on Egocentric Perception, Interaction and Computing (EPIC)

References in corpus (4)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• DeepGaze II: Reading fixations from deep features trained on object recognition
• TurkerGaze: Crowdsourcing Saliency with Webcam based Eye Tracking
• A Deep Spatial Contextual Long-term Recurrent Convolutional Network for Saliency Detection

Cited by in corpus (3)

• Deep Learning for Content-based Personalized Viewport Prediction of 360-Degree VR Videos
• Saliency Prediction on Omnidirectional Images with Generative Adversarial Imitation Learning
• Viewport-Aware Deep Reinforcement Learning Approach for 360${}^o$ Video Caching