Home

Bio

Publications

Research

Courses

Media

Contact

Bootstrapping image classification with sample evaluation


Venkatraman Narayanan, Ruta Desai, Sanjiban Choudhury

Overview


In this project, we design a semi supervised learning method to perform image classification.

Course Details


Course Title: Machine Learning
Instructor: Eric Xing, Aarti Singh

Abstract


In this work, we look at the problem of multiclass image classification in a semisupervised learning framework. Given a small set of labeled images, and a much larger set of unlabeled images, we propose a semisupervised learning method that combines bootstrapping with sample evaluation, to continuously update the learned models for each class. Bootstrapping involves using selflabeled images to retrain the learned models. To overcome the semantic drift that naive bootstrapping is prone to, we use additional sample evaluation methods based on the ideas of cotraining and pairwise constraints, to determine whether or not a newly classified instance should be used for retraining. Experimental results show the usefulness of sample evaluation, when used in conjunction with bootstrapping. In particular, our method is able to achieve a 8% improvement in overall accuracy over baseline bootstrapping, on a 15 class subset of the SUN (Scene UNderstanding) dataset.

Details


Here is the final report.

Here is the final poster.


Links


Here is the link for SUN database and their baseline codes.