Simple vs complex temporal recurrences for video saliency prediction
Published in
British Machine Vision Conference(2019)External Links:
TL
TL;DR
This paper investigates modifying an existing neural network architecture for static saliency prediction using two types of recurrences that integrate information from the temporal domain using a ConvLSTM and a conceptually simple exponential moving average of an internal convolutional state.
Abstract
This paper investigates modifying an existing neural network architecture for static saliency prediction using two types of recurrences that integrate information from the temporal domain. The first modification is the addition of a ConvLSTM within the architecture, while the second is a conceptually simple exponential moving average of an internal convolutional state. We use weights pre-trained on the SALICON dataset and fine-tune our model on DHF1K. Our results show that both modifications achieve state-of-the-art results and produce similar saliency maps. Source code is available at this https URL.
References28 items
1
YouTube-VOS: Sequence-to-Sequence Video Object Segmentation
2
Going from Image to Video Saliency: Augmenting Image Salience with Dynamic Attentional Push
3
Revisiting Video Saliency: A Large-Scale Benchmark and a New Model
4
Excitation Backprop for RNNs
5