Salient Object Detection via Augmented Hypotheses

Introduction

In this work, we propose using augmented hypotheses which consider objectness, foreground

and compactness for salient object detection. Our algorithm consists of four basic steps. First, our method generates the objectness map via objectness hypotheses. Based on the objectness map, we estimate the foreground margin and compute the corresponding foreground map which prefers the foreground objects. From the objectness map and the foreground map, the compactness map is formed to favor the compact objects. We then derive a saliency measure that produces a pixelaccurate saliency map which uniformly covers the objects of interest and consistently separates fore and background.

Figure 1: From top to bottom: original images, the objectness hypotheses, results of our saliency computation, and ground truth labeling. For a better viewing, only 40 object hypotheses are displayed in each image.

Downloads

Paper:

Tam Van Nguyen, Jose Sepulveda: Salient Object Detection via Augmented Hypotheses. IJCAI 2015: 2176-2182 [PDF]

Our salient detection results on MSRA-1000 [Download]

Binary [Download]

How to run binary exe:

- In command prompt, type AH.exe working_folder.

- There must be two sub-folders inside working folder, namely, Images and Saliency

- For example: AH.exe .\Test

----> The source code is available upon request

(Please cite our paper in any published research work making use of this work.)