Skip to main content
ARS Home » Plains Area » Las Cruces, New Mexico » Cotton Ginning Research » Research » Publications at this Location » Publication #74177
Title: EVALUATION OF LEARNING VECTOR QUANTIZATION TO CLASSIFY COTTON TRASH

Author
item Lieberman, Michael
item PATIL, RAJENDRA - LOS ALAMOS NM

Submitted to: Optical Engineering
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 10/22/1996
Publication Date: N/A
Citation: N/A

Interpretive Summary: The cotton industry needs a method to identify the type of trash (nonlint material or (NLM) in cotton samples; this paper evaluates Learning Vector Quantization (LVQ) as that method. LVQ is a classification technique that defines reference vectors (group prototypes) in an N-dimensional feature space. Normalized trash object features extracted from images of compressed cotton samples define the hyperspace. An unknown NLM object is given the label of the closest reference vector (as defined by Euclidean distance). Different feature spaces and NLM classifications are evaluated and accuracies reported for correctly identifying the NLM type. LVQ is used to partition cotton trash into: (1) bark (B), leaf (L), pepper (P), or stick (S); (2) bark, combined leaf and pepper (LP), or stick; or (3) bark and nonbark (N). Percent accuracy for correctly identifying 139 pieces of test trash placed on laboratory prepared samples for the three scenarios is reported.

Technical Abstract: The cotton industry needs a method to identify the type of trash (nonlint material or (NLM) in cotton samples; this paper evaluates Learning Vector Quantization (LVQ) as that method. LVQ is a classification technique that defines reference vectors (group prototypes) in an N-dimensional feature space. An internet source for the source code is given in the references. Normalized trash object features extracted from images of compressed cotto samples define the hyperspace. An unknown NLM object is given the label of the closest reference vector (as defined by Euclidean distance). Different normalized feature spaces and NLM classifications are evaluated and accuracies reported for correctly identifying the NLM type. LVQ is used to partition cotton trash into: (1) bark (B), leaf (L), pepper (P), or stick (S); (2) bark, combined leaf and pepper (LP), or stick; or (3) bark and nonbark (N). Percent accuracy for correctly identifying 139 pieces of test ttrash placed on laboratory prepared samples for the three scenarios are {B 95, L: 87, P: 100, S: 88}, {B: 95, LP: 99, S: 88}, and {B: 100, N: 97}, respectively. Also, LVQ results are compared to previous work using backpropagating neural networks.