Oregon State University



Event Details

MS Final Examination – Chenyu Wang

Tuesday, February 28, 2017 10:00 AM - 12:00 PM

Fine Grained Video Classification for Endangered Bird Species Protection
This project is about recognizing and detecting eagles versus other similar bird species in in low-resolution videos. It is a part of a larger research project aimed at developing an environmentally safe wind-turbine energy systems. Our contributions on this project include: (1) Developing a deep recurrent neural architecture for eagle detection, and (2) Compiling a new dataset of videos for training and testing our approach. Our approach uses a deep convolutional neural network (CNN) to extract deep features from each input video frame, and a long-short term memory network (LSTM) for fusing the CNN’s outputs and predicting the bird species that occurs in the video. We studied several variants of this architecture, under different settings that account for optical flow as additional motion cues. Our results demonstrate a relatively high accuracy of eagle recognition versus similar bird species in very challenging videos.

Major Advisor: Sinisa Todorovic
Committee: Eugene Zhang
Committee: Lizhong Cheng

Kelley Engineering Center (campus map)
Nicole Thompson
1 541 737 3617
Nicole.Thompson at oregonstate.edu
Sch Elect Engr/Comp Sci
This event appears on the following calendars: