Unsupervised Deep Feature Transfer for Low Resolution Image Classification

TL;DR

This paper introduces an unsupervised deep feature transfer method that enhances low-resolution image classification by leveraging pre-trained CNN features and a simple transfer network, improving accuracy without fine-tuning.

Contribution

The proposed method transfers high-resolution features to low-resolution images without fine-tuning CNNs, serving as a plug-in module for improved low-res image classification.

Findings

Significant accuracy improvements over baseline methods.

Effective feature transfer preserves data structure.

Compatible with state-of-the-art deep networks.

Abstract

In this paper, we propose a simple while effective unsupervised deep feature transfer algorithm for low resolution image classification. No fine-tuning on convenet filters is required in our method. We use pre-trained convenet to extract features for both high- and low-resolution images, and then feed them into a two-layer feature transfer network for knowledge transfer. A SVM classifier is learned directly using these transferred low resolution features. Our network can be embedded into the state-of-the-art deep neural networks as a plug-in feature enhancement module. It preserves data structures in feature space for high resolution images, and transfers the distinguishing features from a well-structured source domain (high resolution features space) to a not well-organized target domain (low resolution features space). Extensive experiments on VOC2007 test set show that the proposed method achieves significant improvements over the baseline of using feature extraction.