A Compact Deep Architecture for Real-time Saliency Prediction

TL;DR

This paper introduces a compact, efficient deep neural network architecture for real-time saliency prediction, combining novel layers and modifications to existing models to achieve high accuracy and speed suitable for edge devices.

Contribution

The paper presents a new lightweight deep learning model with a modified U-net, a novel fully connected layer, and central difference convolutional layers for improved real-time saliency prediction.

Findings

Outperforms state-of-the-art models on benchmark datasets

Achieves real-time processing speeds

Balances accuracy with computational efficiency

Abstract

Saliency computation models aim to imitate the attention mechanism in the human visual system. The application of deep neural networks for saliency prediction has led to a drastic improvement over the last few years. However, deep models have a high number of parameters which makes them less suitable for real-time applications. Here we propose a compact yet fast model for real-time saliency prediction. Our proposed model consists of a modified U-net architecture, a novel fully connected layer, and central difference convolutional layers. The modified U-Net architecture promotes compactness and efficiency. The novel fully-connected layer facilitates the implicit capturing of the location-dependent information. Using the central difference convolutional layers at different scales enables capturing more robust and biologically motivated features. We compare our model with state of the art saliency models using traditional saliency scores as well as our newly devised scheme. Experimental results over four challenging saliency benchmark datasets demonstrate the effectiveness of our approach in striking a balance between accuracy and speed. Our model can be run in real-time which makes it appealing for edge devices and video processing.