Mr. Gefei Yang Profile

Gefei Yang

at Rochester Institute of Technology

SPIE Involvement:

Author

Publications (2)

This will count as one of your downloads.

You will have access to both the presentation and article (if available).

DOWNLOAD NOW

This content is available for download via your institution's subscription. To access this item, please sign in to your personal account.

Email or Username Forgot your username?

Password Forgot your password?

Show

Keep me signed in

No SPIE account? Create an account

Proceedings Article | 8 May 2018 Paper

A wide dense architecture neural network for material classification based on spectral BTF

Gefei Yang, Michael Gartley

Proceedings Volume 10644, 106441R (2018) https://doi.org/10.1117/12.2302911

KEYWORDS: Neural networks, Data modeling, Feature extraction, Image classification, Convolution, Network architectures, Reflectivity, Light sources and illumination, Performance modeling, Bidirectional reflectance transmission function

Read Abstract +

Traditionally, material surface classification has relied on reflectance spectrum measurement and pixel-wise comparison. For best results, the measurement usually requires measurement of the full spectral reflectance which can be the time-consuming issue and error-prone. Our previous works have proved that convolutional neural networks could learn the Bi-directional Texture Function (BTF) data feature hierarchy from pixels to the classifier. In this paper, we trained a wide-dense neural network for BTF data. The dense neural network was developed from the residual network (ResNet), which made a residual mapping for these stacked layers instead of directly fitting a desired underlying mapping. This shortcut idea has excellent performance in deeper layers. The dense structure reduced the redundancy within the feature maps of the individual layers and increased training speed. The narrow size of the dense net layers simplified the feature map for the whole network. The wide structure of the network was applied due to the resolution of our BTF data. In this paper, we generated BTF angular maps for material classification. Our data consisted of 151 lighting and 151 viewing directions. This angular resolution is higher than input data of conventional pre-trained dense network model. By adding a wide net structure, training time is reduced and classification performance improved. Training our own networks for specific data requires a large training dataset, therefore we augmented angular BTF image to improve the robustness of the training model. Within the over 30,000 BTF angular maps, we obtained a more reliable training model. Finally, we compared the improvement of our wide-dense network with other pre-trained neural networks and other feature extraction network.

Proceedings Article | 5 May 2017 Presentation + Paper

Application of a neural network for reflectance spectrum classification

Gefei Yang, Michael Gartley

Proceedings Volume 10198, 1019817 (2017) https://doi.org/10.1117/12.2261847

KEYWORDS: Neural networks, Reflectivity, Bidirectional reflectance transmission function, Machine learning, Convolution, Image analysis, Light sources and illumination, Cameras, Feature extraction, Image classification, Databases, Data modeling, Image processing, Computer simulations

Read Abstract +

Traditional reflectance spectrum classification algorithms are based on comparing spectrum across the electromagnetic spectrum anywhere from the ultra-violet to the thermal infrared regions. These methods analyze reflectance on a pixel by pixel basis. Inspired by high performance that Convolution Neural Networks (CNN) have demonstrated in image classification, we applied a neural network to analyze directional reflectance pattern images. By using the bidirectional reflectance distribution function (BRDF) data, we can reformulate the 4-dimensional into 2 dimensions, namely incident direction × reflected direction × channels. Meanwhile, RIT’s micro-DIRSIG model is utilized to simulate additional training samples for improving the robustness of the neural networks training. Unlike traditional classification by using hand-designed feature extraction with a trainable classifier, neural networks create several layers to learn a feature hierarchy from pixels to classifier and all layers are trained jointly. Hence, the our approach of utilizing the angular features are different to traditional methods utilizing spatial features. Although training processing typically has a large computational cost, simple classifiers work well when subsequently using neural network generated features. Currently, most popular neural networks such as VGG, GoogLeNet and AlexNet are trained based on RGB spatial image data. Our approach aims to build a directional reflectance spectrum based neural network to help us to understand from another perspective. At the end of this paper, we compare the difference among several classifiers and analyze the trade-off among neural networks parameters.

My Library

You currently do not have any folders to save your paper to! Create a new folder below.

Folder Name

Folder Description

View contact details

UPDATE YOUR PROFILE

Is this your profile? Update it now.

Sign into your SPIE.org account

Don’t have a profile and want one?

Create an account on SPIE.org

Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks. You are receiving this notice because your organization may not have SPIE eBooks access.*

*Shibboleth/Open Athens users─please sign in to access your institution's subscriptions.

To obtain this item, you may purchase the complete book in print or electronic format on SPIE.org.

ORGANIZATIONAL
Sign in with credentials provided by your organization.

Organizational Username

Organizational Password

Show/Hide Password

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available

Members:

Non-members: ADD TO CART

Keywords/Phrases

Search In:

Publication Years