Please use this identifier to cite or link to this item:
https://hdl.handle.net/11681/37618Full metadata record
| DC Field | Value | Language |
|---|---|---|
| dc.contributor.author | Donovan, Jordan T. | en_US |
| dc.creator | Information Technology Laboratory (U.S.) | en_US |
| dc.date.accessioned | 2020-07-30T17:08:41Z | en_US |
| dc.date.available | 2020-07-30T17:08:41Z | en_US |
| dc.date.issued | 2020-07 | en_US |
| dc.identifier.govdoc | ERDC/ITL MP-20-1 | en_US |
| dc.identifier.uri | https://hdl.handle.net/11681/37618 | en_US |
| dc.identifier.uri | http://dx.doi.org/10.21079/11681/37618 | en_US |
| dc.description | Miscellaneous Paper | en_US |
| dc.description.abstract | Neural networks (NNs) excel at solving several complex, non-linear problems in the area of supervised learning. A prominent application of these networks is image classification. Numerous improvements over the last few decades have improved the capability of these image classifiers. However, neural networks are still a black-box for solving image classification and other sophisticated tasks. A number of experiments conducted look into exactly how neural networks solve these complex problems. This paper dismantles the neural network solution, incorporating convolution layers, of a specific material classifier. Several techniques are utilized to investigate the solution to this problem. These techniques look at specifically which pixels contribute to the decision made by the NN as well as a look at each neuron’s contribution to the decision. The purpose of this investigation is to understand the decision-making process of the NN and to use this knowledge to suggest improvements to the material classification algorithm. | en_US |
| dc.description.sponsorship | Engineer Research and Development Center (U.S.) | en_US |
| dc.description.tableofcontents | ABSTRACT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii LIST OF TABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vi LIST OF FIGURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii CHAPTER 1. INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.1 Neural Networks and Visualization . . . . . . . . . . . . . . . . 1 1.2 Material Classification . . . . . . . . . . . . . . . . . . . . . . . 3 1.3 Contribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 2. LITERATURE REVIEW . . . . . . . . . . . . . . . . . . . . . . . . . . 5 2.1 Neural Network Visualization . . . . . . . . . . . . . . . . . . . 5 2.2 Material Classification . . . . . . . . . . . . . . . . . . . . . . . 11 2.2.1 Material Databases . . . . . . . . . . . . . . . . . . . . . 11 2.2.2 Material Recognition . . . . . . . . . . . . . . . . . . . . 14 2.2.3 Convolutional Neural Networks (CNNs) . . . . . . . . . . 15 3. METHODOLOGY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 3.1 Hypothesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 3.2 Contribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 3.3 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 4. EXPERIMENT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 4.1 Material Classifier Replication . . . . . . . . . . . . . . . . . . . 21 4.2 Neural Network Visualization . . . . . . . . . . . . . . . . . . . 24 4.3 In-depth Comparison of Material Classification Methods and Recommendations for Improvement . . . . . . . . . . . . . . . . . . 26 5. RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 5.1 Material Classifier Analysis . . . . . . . . . . . . . . . . . . . . 27 5.2 Visualization Tool Analysis . . . . . . . . . . . . . . . . . . . . 40 5.3 In-Depth Material Classification Technique Comparison . . . . . 45 5.3.1 Material Databases . . . . . . . . . . . . . . . . . . . . . 45 5.3.2 Material Features and Recognition Techniques . . . . . . . 54 5.3.3 Convolutional Neural Networks . . . . . . . . . . . . . . . 71 5.4 Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . 73 6. DISCUSSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 7. CONCLUSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 | en_US |
| dc.format.extent | 100 pages / 27.53 MB | en_US |
| dc.format.medium | PDF/A | en_US |
| dc.language.iso | en_US | en_US |
| dc.publisher | Engineer Research and Development Center (U.S.) | en_US |
| dc.relation.ispartofseries | Miscellaneous Paper (Information Technology Laboratory (U.S.)) ; no. ERDC/ITL MP-20-1 | en_US |
| dc.rights | Approved for Public Release; Distribution is Unlimited | en_US |
| dc.source | This Digital Resource was created in Microsoft Word and Adobe Acrobat | en_US |
| dc.subject | Neural networks (Computer science) | en_US |
| dc.subject | Machine learning | en_US |
| dc.subject | Computer vision | en_US |
| dc.subject | Image processing | en_US |
| dc.subject | Digital images--Deconvolution | en_US |
| dc.subject | Deep visualization | en_US |
| dc.subject | Image segmentation | en_US |
| dc.subject | Object detection | en_US |
| dc.subject | Material classification | en_US |
| dc.title | Understanding state-of-the-art material classification through deep visualization | en_US |
| dc.type | Report | en_US |
| Appears in Collections: | Miscellaneous Paper | |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| ERDC-ITL MP-20-1.pdf | ERDC/ITL MP-20-1 | 27.53 MB | Adobe PDF |  View/Open |