A proposal of prior probability-oriented clustering in feature encoding strategies.
Codebook-based feature encodings are a standard framework for image recognition issues. A codebook is usually constructed by clusterings, such as the k-means and the Gaussian Mixture Model (GMM). A codebook size is an important factor to decide the trade-off between recognition performance and compu...
Main Authors: | Yuki Shinomiya, Yukinobu Hoshino |
---|---|
Format: | Article |
Language: | English |
Published: |
Public Library of Science (PLoS)
2019-01-01
|
Series: | PLoS ONE |
Online Access: | https://doi.org/10.1371/journal.pone.0210146 |
Similar Items
-
Processing of prior probability
by: Scheibe, Christina
Published: (2010) -
The Role of Prior Probability in Forensic Assessments
by: William C. Thompson, et al.
Published: (2013-10-01) -
Noticing relevant problem features: Activating prior knowledge affects problem solving by guiding encoding
by: Noelle M Crooks, et al.
Published: (2013-11-01) -
Human online adaptation to changes in prior probability.
by: Elyse H Norton, et al.
Published: (2019-07-01) -
Prior probabilities and thermal characteristics of heat engines
by: Aneja Preety, et al.
Published: (2012-06-01)