Sponsor
Portland State University. Department of Computer Science
First Advisor
Melanie Mitchell
Date of Publication
Spring 6-12-2018
Document Type
Thesis
Degree Name
Master of Science (M.S.) in Computer Science
Department
Computer Science
Subjects
Reinforcement learning, Machine learning, Computer vision
DOI
10.15760/etd.6322
Physical Description
1 online resource (vii, 46 pages)
Abstract
In this thesis, I explore a reinforcement learning technique for improving bounding box localizations of objects in images. The model takes as input a bounding box already known to overlap an object and aims to improve the fit of the box through a series of transformations that shift the location of the box by translation, or change its size or aspect ratio. Over the course of these actions, the model adapts to new information extracted from the image. This active localization approach contrasts with existing bounding-box regression methods, which extract information from the image only once. I implement, train, and test this reinforcement learning model using data taken from the Portland State Dog-Walking image set.
The model balances exploration with exploitation in training using an ε-greedy policy. I find that the performance of the model is sensitive to the ε-greedy configuration used during training, performing best when the epsilon parameter is set to very low values over the course of training. With = 0.01, I find the algorithm can improve bounding boxes in about 78% of test cases for the "dog" object category, and 76% for the "human" category.
Persistent Identifier
https://archives.pdx.edu/ds/psu/25617
Recommended Citation
Cleland, Andrew Lewis, "Bounding Box Improvement with Reinforcement Learning" (2018). Dissertations and Theses. Paper 4438.
https://doi.org/10.15760/etd.6322