Evaluating Deep Convolutional Neural Networks for Material Classification

TL;DR

This paper empirically evaluates various state-of-the-art CNN architectures for material classification from images, demonstrating high accuracy on challenging datasets and providing insights into their comparative performance.

Contribution

It offers a comprehensive comparison of CNN architectures for material classification, highlighting their effectiveness and identifying the best-performing models.

Findings

CNNs can achieve up to 94.99% mean average precision in material classification.

Different CNN architectures vary significantly in performance.

The study provides benchmarks for future research in material recognition.

Abstract

Determining the material category of a surface from an image is a demanding task in perception that is drawing increasing attention. Following the recent remarkable results achieved for image classification and object detection utilising Convolutional Neural Networks (CNNs), we empirically study material classification of everyday objects employing these techniques. More specifically, we conduct a rigorous evaluation of how state-of-the art CNN architectures compare on a common ground over widely used material databases. Experimental results on three challenging material databases show that the best performing CNN architectures can achieve up to 94.99\% mean average precision when classifying materials.