CRL: Class Representative Learning for Image Classification

TL;DR

The paper introduces CRL, a novel class representative learning model for image classification that improves recognition performance by aggregating features into class representatives, especially effective in zero-shot learning scenarios.

Contribution

CRL is a new model that constructs class representatives from CNN features and matches them with new data, enhancing zero-shot and mobile deep learning performance.

Findings

Outperforms state-of-the-art in ZSL and mobile deep learning

Demonstrates significant accuracy improvements on benchmark datasets

Efficient in distributed environments using Apache Spark

Abstract

Building robust and real-time classifiers with diverse datasets are one of the most significant challenges to deep learning researchers. It is because there is a considerable gap between a model built with training (seen) data and real (unseen) data in applications. Recent works including Zero-Shot Learning (ZSL), have attempted to deal with this problem of overcoming the apparent gap through transfer learning. In this paper, we propose a novel model, called Class Representative Learning Model (CRL), that can be especially effective in image classification influenced by ZSL. In the CRL model, first, the learning step is to build class representatives to represent classes in datasets by aggregating prominent features extracted from a Convolutional Neural Network (CNN). Second, the inferencing step in CRL is to match between the class representatives and new data. The proposed CRL model demonstrated superior performance compared to the current state-of-the-art research in ZSL and mobile deep learning. The proposed CRL model has been implemented and evaluated in a parallel environment, using Apache Spark, for both distributed learning and recognition. An extensive experimental study on the benchmark datasets, ImageNet-1K, CalTech-101, CalTech-256, CIFAR-100, shows that CRL can build a class distribution model with drastic improvement in learning and recognition performance without sacrificing accuracy compared to the state-of-the-art performances in image classification.