Convolutional Extreme Learning Machines: A Systematic Review

Iago Richard Rodrigues,Patricia Takako Endo,Judith Kelner,Sebastião Rogério Da Silva Neto,Djamel Sadok

doi:10.3390/informatics8020033

Iago Richard Rodrigues, Patricia Takako Endo + Show 3 more

Open Access

https://doi.org/10.3390/informatics8020033

Copy DOI

Abstract

Much work has recently identified the need to combine deep learning with extreme learning in order to strike a performance balance with accuracy, especially in the domain of multimedia applications. When considering this new paradigm—namely, the convolutional extreme learning machine (CELM)—we present a systematic review that investigates alternative deep learning architectures that use the extreme learning machine (ELM) for faster training to solve problems that are based on image analysis. We detail each of the architectures that are found in the literature along with their application scenarios, benchmark datasets, main results, and advantages, and then present the open challenges for CELM. We followed a well-structured methodology and established relevant research questions that guided our findings. Based on 81 primary studies, we found that object recognition is the most common problem that is solved by CELM, and CCN with predefined kernels is the most common CELM architecture proposed in the literature. The results from experiments show that CELM models present good precision, convergence, and computational performance, and they are able to decrease the total processing time that is required by the learning process. The results presented in this systematic review are expected to contribute to the research area of CELM, providing a good starting point for dealing with some of the current problems in the analysis of computer vision based on images.

Highlights

Because of the growth of image analysis-based applications, researchers have adopted deep learning to develop intelligent systems that provide learning tasks in computer vision, image processing, text recognition, and other signal processing problems
Based on the scenarios and the most common datasets used in the primary studies, in this subsection we describe the main findings when applying the convolutional extreme learning machine (CELM) to image analysis
We present the time that is required for training and testing the CELM models

Summary

Introduction

Because of the growth of image analysis-based applications, researchers have adopted deep learning to develop intelligent systems that provide learning tasks in computer vision, image processing, text recognition, and other signal processing problems. Unlike classic approaches to pattern recognition tasks, convolutional neural networks (CNNs), a type of deep learning, can perform the process of extracting features and, at the same time, recognize these features. CNNs can process data that are stored as multidimensional arrays (1D, 2D, and so on). They extract meaningful abstract representations from raw data [1], such as images, audio, text, video, and so on. CNNs have received attention in the last decade due to their success in fields such as image classification [2], object detection [3], semantic segmentation [4], and medical applications that support a diagnosis by signals or images [5]

Objectives

Methods

Results

Conclusion