Deep Learning of Image Representations with Convolutional Neural Networks Autoencoder for Image Retrieval with Relevance Feedback

Abstract

mage retrieval with traditional relevance feedback encounters problems: (1) ability to represent handcrafted features which is limited, and (2) inefficient with
high-dimensional data such as image data. In this paper,
we propose a framework based on very deep convolutional
neural network autoencoder for image retrieval, called AIR
(Autoencoders for Image Retrieval). Our proposed framework
allows to learn feature vectors directly from the raw image
and in an unsupervised manner. In addition, our framework
utilizes a hybrid approach of unsupervised and supervised
to improve retrieval performance. The experimental results
show that our method gives better results than some existing
methods on the CIFAR-100 image set, which consists of 60,000
images.