Sciweavers

5 search results - page 1 / 1
» Dynamically Weighted Hidden Markov Model for Spam Deobfuscat...
Sort
View
IJCAI
2007
13 years 6 months ago
Dynamically Weighted Hidden Markov Model for Spam Deobfuscation
Spam deobfuscation is a processing to detect obfuscated words appeared in spam emails and to convert them back to the original words for correct recognition. Lexicon tree hidden M...
Seunghak Lee, Iryoung Jeong, Seungjin Choi
CEAS
2005
Springer
13 years 10 months ago
Spam Deobfuscation using a Hidden Markov Model
To circumvent spam filters, many spammers attempt to obfuscate their emails by deliberately misspelling words or introducing other errors into the text. For example viagra may be...
Honglak Lee, Andrew Y. Ng
ICRA
1998
IEEE
141views Robotics» more  ICRA 1998»
13 years 9 months ago
On Discontinuous Human Control Strategies
Models of human control strategy (HCS), which accurately emulate dynamic human behavior, have far reaching potential in areas ranging from robotics to virtual reality to the intel...
Michael C. Nechyba, Yangsheng Xu
CVPR
2009
IEEE
15 years 3 days ago
A Similarity Measure Between Vector Sequences with Application to Handwritten Word Image Retrieval
This article proposes a novel similarity measure between vector sequences. Recently, a model-based approach was introduced to address this issue. It consists in modeling each se...
José A. Rodríguez-Serrano, Florent P...
MVA
2007
207views Computer Vision» more  MVA 2007»
13 years 6 months ago
View-invariant Human Action Recognition Based on Factorization and HMMs
of the fundamental challenges of human action recognition is accounting for the variability that arises during video capturing. For a specific action class, the 2D observations of...
Xi Li, Kazuhiro Fukui