CVFeb 26, 2019

Object Discovery From a Single Unlabeled Image by Mining Frequent Itemset With Multi-scale Features

arXiv:1902.09968v334 citations
Originality Incremental advance
AI Analysis

This addresses the problem of unsupervised object discovery for computer vision applications, but it is incremental as it builds on existing pattern mining and pre-trained CNN features.

The paper tackles the problem of discovering dominant objects from a single unlabeled image, which is more challenging than typical localization tasks, and proposes Object Location Mining (OLM), achieving competitive localization performance compared to state-of-the-art methods on various benchmarks.

TThe goal of our work is to discover dominant objects in a very general setting where only a single unlabeled image is given. This is far more challenge than typical co-localization or weakly-supervised localization tasks. To tackle this problem, we propose a simple but effective pattern mining-based method, called Object Location Mining (OLM), which exploits the advantages of data mining and feature representation of pre-trained convolutional neural networks (CNNs). Specifically, we first convert the feature maps from a pre-trained CNN model into a set of transactions, and then discovers frequent patterns from transaction database through pattern mining techniques. We observe that those discovered patterns, i.e., co-occurrence highlighted regions, typically hold appearance and spatial consistency. Motivated by this observation, we can easily discover and localize possible objects by merging relevant meaningful patterns. Extensive experiments on a variety of benchmarks demonstrate that OLM achieves competitive localization performance compared with the state-of-the-art methods. We also evaluate our approach compared with unsupervised saliency detection methods and achieves competitive results on seven benchmark datasets. Moreover, we conduct experiments on fine-grained classification to show that our proposed method can locate the entire object and parts accurately, which can benefit to improving the classification results significantly.

Code Implementations1 repo
Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes