Modified UDP-Based semi-supervised learning for fruit internal quality detection

Authors:
Peiyi Zhu;Benlian Xu;Jue Gao
Affiliations:
School of Electrical and Automation Engineering, Changshu Institute of technology, Changshu, Jiangsu, China;School of Electrical and Automation Engineering, Changshu Institute of technology, Changshu, Jiangsu, China;School of Electrical and Automation Engineering, Changshu Institute of technology, Changshu, Jiangsu, China
Venue:
ICSI'12 Proceedings of the Third international conference on Advances in Swarm Intelligence - Volume Part II
Year:
2012

Citing 3
Cited 0

Laplacian Eigenmaps for dimensionality reduction and data representation

Neural Computation
Globally Maximizing, Locally Minimizing: Unsupervised Discriminant Projection with Applications to Face and Palm Biometrics

IEEE Transactions on Pattern Analysis and Machine Intelligence
Collision detection between complex polyhedra

Computers and Graphics

Quantified Score

Hi-index	0.00

Visualization

Abstract

In order to ignore the limitation of local structure features for the traditional linear dimensionality reduction methods, a new semi-supervised manifold learning is proposed for apple mealiness detection. Assuming the character of the hyperspectral scattering images, an unsupervised non-linear dimensionality reduction method unsupervised discriminant projection (UDP) coupled with sample label information and then develop a semi-supervised learning algorithm, which can keep the local and global structure and can take advantage of the important label information, then get geometric structure optimal linear projection. The classification results with PCA-MUDP are compared with some traditional algorithm. To the two-class classification of ‘mealy' and ‘non-mealy' apples, the results show that PCA-MUDP is better than the others.