SalProp: Salient object proposals via aggregated edge cues

TL;DR

This paper introduces SalProp, a novel object proposal method that leverages a graph-based salient edge classification framework and Bayesian edge maps to improve object detection efficiency and accuracy.

Contribution

It presents a new edge-based object proposal generation approach using probabilistic edge maps and CRF for classification, achieving competitive results with fewer proposals.

Findings

Competitive performance on PASCAL VOC 2007

Fewer proposals needed for effective detection

Outperforms 10 popular detection techniques

Abstract

In this paper, we propose a novel object proposal generation scheme by formulating a graph-based salient edge classification framework that utilizes the edge context. In the proposed method, we construct a Bayesian probabilistic edge map to assign a saliency value to the edgelets by exploiting low level edge features. A Conditional Random Field is then learned to effectively combine these features for edge classification with object/non-object label. We propose an objectness score for the generated windows by analyzing the salient edge density inside the bounding box. Extensive experiments on PASCAL VOC 2007 dataset demonstrate that the proposed method gives competitive performance against 10 popular generic object detection techniques while using fewer number of proposals.