Files in this item

FilesDescriptionFormat

application/pdf

application/pdfECE499-Sp2018-du.pdf (1MB)
(no description provided)PDF

Description

Title:Visualizing and interpreting a deep neural network for classification of vehicular orientation
Author(s):Du, Jianlin
Contributor(s):Varshney, Lav R.
Subject(s):Deep learning
Abstract:We trained a deep neural network to classify images of cars facing 36 different directions, on a 2D image dataset rendered from 3D car models. After achieving a validation accuracy of 98.23%, we applied a series of interpretation techniques, including semantic dictionary, spatial attribution, and channel attribution, to the trained model, which enable us gain important insights on how the model recognized a car’s direction. For example, the channel attribution technique reveals in a certain layer, which filters contribute the most to distinguish a front facing car from a right facing car. Moreover, there are interactive interfaces for all the experiments, and readers could explore the interpretation of the model in a notebook environment.
Issue Date:2018-05
Genre:Other
Type:Text
Language:English
URI:http://hdl.handle.net/2142/99989
Date Available in IDEALS:2018-05-23


This item appears in the following Collection(s)

Item Statistics