Retinomorphic Event-Driven Representations for Video Tasks

Simple item page
dc.contributor.advisorShrivastava, Anshumalien_US
dc.contributor.committeeMemberPatel, Ankit B.en_US
dc.creatorLiu, Wanjiaen_US
dc.date.accessioned2019-05-17T13:29:15Zen_US
dc.date.available2019-05-17T13:29:15Zen_US
dc.date.created2017-12en_US
dc.date.issued2017-11-30en_US
dc.date.submittedDecember 2017en_US
dc.date.updated2019-05-17T13:29:15Zen_US
dc.description.abstractDeep neural networks have revolutionized static image understanding through their amazing performance in tasks such as classification, segmentation and style transfer. However, current architectures have yet to find as much success in video tasks mainly due to increased data dimension, higher information throughput and a sub-optimal frame-driven representation. Inspired by the early vision systems of mammals and insects, we propose an event-driven input representation (EDR) that models several major properties of early retinal circuits: 1. Output scales logrithmically with input intensity, 2. ON/OFF pathways for differentiating event types, 3. Relative change detection and event detection via thresholding. With UCF-101 video action recognition experiments, we show that neural network utilizing EDR as input performs near state-of-the-art in accuracy while achieving a 1,500x speedup in input representation processing (9K frames/sec), as compared to optical flow. EDR's fast processing speed enables real-time inference/learning in time sensitive video applications such as reinforcement learning. In this vein, we use EDR to demonstrate performance improvements over state-of-the-art reinforcement learning algorithms for Atari games, something that has not been possible with optical flow.en_US
dc.format.mimetypeapplication/pdfen_US
dc.identifier.citationLiu, Wanjia. "Retinomorphic Event-Driven Representations for Video Tasks." (2017) Master’s Thesis, Rice University. <a href="https://hdl.handle.net/1911/105601">https://hdl.handle.net/1911/105601</a>.en_US
dc.identifier.urihttps://hdl.handle.net/1911/105601en_US
dc.language.isoengen_US
dc.rightsCopyright is held by the author, unless otherwise indicated. Permission to reuse, publish, or reproduce the work beyond the bounds of fair use or other exemptions to copyright law must be obtained from the copyright holder.en_US
dc.subjectRetinomorphicen_US
dc.subjectEvent-Drivenen_US
dc.subjectMotion Representationen_US
dc.subjectOptical-Flowen_US
dc.subjectAction Recognitionen_US
dc.subjectReinforcement Learningen_US
dc.titleRetinomorphic Event-Driven Representations for Video Tasksen_US
dc.typeThesisen_US
dc.type.materialTexten_US
thesis.degree.departmentComputer Scienceen_US
thesis.degree.disciplineEngineeringen_US
thesis.degree.grantorRice Universityen_US
thesis.degree.levelMastersen_US
thesis.degree.majorDeep Learningen_US
thesis.degree.nameMaster of Scienceen_US
Files
Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
LIU-DOCUMENT-2017.pdf
Size:
3.55 MB
Format:
Adobe Portable Document Format
Download
License bundle
Now showing 1 - 2 of 2
No Thumbnail Available
Name:
PROQUEST_LICENSE.txt
Size:
5.84 KB
Format:
Plain Text
Description:
Download
No Thumbnail Available
Name:
LICENSE.txt
Size:
2.6 KB
Format:
Plain Text
Description:
Download
Collections
Rice University Theses and Dissertations