Proceedings of the Annual Conference of Biomedical Fuzzy Systems Association
Online ISSN : 2424-2586
Print ISSN : 1345-1510
ISSN-L : 1345-1510
24
Session ID : 29P-C-4
Conference information

29P-C-4 Consideration on Effectiveness of Pseudo Training Data in CHLAC Features-based Crows' Movement Detection
Kouji OCHIAITakanori KOGANoriaki SUETAKEEiji UCHINO
Author information
CONFERENCE PROCEEDINGS FREE ACCESS

Details
Abstract
Ducks fed in rice paddies for rice-duck farming are preyed by injurious birds such as crows. The aims of this study are to detect crows' predation-motion by using time-varying image processing technique, and to prevent the crows' predation. CHLAC feature is useful to detect crows' motions. However, when the target is a wild bird such as a crow, it is difficult to correct large amounts of training data. In this report, we propose the way to improve recognition rate by using pseudo data.
Content from these authors
© 2011 Biomedical Fuzzy Systems Association
Previous article Next article
feedback
Top