Concept Map: Mining noisy web data for concept learning


Work by --

Eren Golge
and Pinar Duygulu

Related Paper --

Golge, E., & Duygulu, P.. ConceptMap:Mining noisy web data for concept learning , The European Conference on Computer Vision (ECCV) 2014.

Abstract --

We attack the problem of learning concepts automatically from noisy Web image search results. The idea is based on discovering common characteristics shared among subsets of images by posing a method that is able to organise the data while eliminating irrelevant instances. We propose a novel clustering and outlier detection method, namely Concept Map (CMAP). Given an image collection returned for a concept query, CMAP provides clusters pruned from outliers. Each cluster is used to train a model representing a different characteristics of the concept. The proposed method outperforms the state-of-the-art studies on the task of learning from noisy web data for low-level attributes, as well as high level object categories. It is also competitive with the supervised methods in learning scene concepts. Moreover, results on naming faces support the generalisation capability of the CMAP framework to different domains. CMAP is capable to work at large scale with no supervision through exploiting the available sources

 

Image and video hosting by TinyPic
CMAP in runtime. It shows the unit updates from iteration 1 to saturation where the units are precisely defined.

 

 

System Pipeline

 

Visual Examples --

 

ConceptMap attribute clusters and outliers.

 

ConceptMap face results.

 

Tasks and Datasets --

We give compelling experimental results for the different tasks; Notice: Our datasets that we used to learn concept models in the work will be released soon !!

Codes --

The code for clustering and outlier detection (RSOM) is in GITHUB !!

Results --

Attribute classification results in various state of art datasets.

 

Scene classification results in Scene-15 and MIT-indoor.

 

Object classification results in the dataset released by Fergus et al. .

 

Face classification results for FAN-large.