Segmentation Free Object Discovery in Video

TL;DR

This paper introduces a simple, unsupervised method for extending static image object proposals to videos by exploiting spatial correlations over time, enabling object discovery without annotations.

Contribution

It presents a novel, supervision-free approach to generate spatio-temporal object tracks in videos and proposes a new dataset-independent evaluation method based on classifier entropy.

Findings

Effective in generating object tracks with minimal supervision

Works well on YouTube Objects and ILSVRC-2015 VID datasets

Provides a new evaluation metric for object proposals

Abstract

In this paper we present a simple yet effective approach to extend without supervision any object proposal from static images to videos. Unlike previous methods, these spatio-temporal proposals, to which we refer as tracks, are generated relying on little or no visual content by only exploiting bounding boxes spatial correlations through time. The tracks that we obtain are likely to represent objects and are a general-purpose tool to represent meaningful video content for a wide variety of tasks. For unannotated videos, tracks can be used to discover content without any supervision. As further contribution we also propose a novel and dataset-independent method to evaluate a generic object proposal based on the entropy of a classifier output response. We experiment on two competitive datasets, namely YouTube Objects and ILSVRC-2015 VID.