Arboreal Identification Supported by Fuzzy Modeling for Trunk Texture Recognition

BRESSANE, A.; FENGLER, F.H.; ROVEDA, S.R.M.M.; ROVEDA, J.A.F.; MARTINS, A.C.G.

doi:10.5540/tema.2018.019.01.0111

Acessibilidade / Reportar erro

Brasil

Español English

sumário « anterior atual seguinte »

Sumário

Articles • TEMA (São Carlos) 19 (1) • Jan-Apr 2018 • https://doi.org/10.5540/tema.2018.019.01.0111 copy

Arboreal Identification Supported by Fuzzy Modeling for Trunk Texture Recognition

Authorship SCIMAGO INSTITUTIONS RANKINGS

ABSTRACT

Due to the natural variability of the arboreal bark there are texture patterns in trunk images with values belonging to more than one species. Thus, the present study analyzed the usage of fuzzy modeling as an alternative to handle the uncertainty in the trunk texture recognition, in comparison with other machine learning algorithms. A total of 2160 samples, belonging to 20 tree species from the Brazilian native deciduous forest, were used in the experimental analyzes. After transforming the images from RGB to HSV, 70 texture patterns have been extracted based on first and second order statistics. Secondly, an exploratory factor analysis was performed for dealing with redundant information and optimizing the computational effort. Then, only the first dimensions with higher cumulative variability were selected as input variables in the predictive modeling. As a result, fuzzy modeling reached a generalization ability that outperformed some algorithms widely used in classification tasks. Therefore, the fuzzy modeling can be considered as a competitive approach, with reliable performance in arboreal trunk texture recognition.

Keywords:
soft computing; image processing; pattern matching; bioinformatics

Learning algorithm	Control parameters settings	Available in
Fuzzy-Based System (FRBCS)	frbcs.w, mf: gaussian, t-norm: product, antecedent terms: 23	Package ‘frbs’ R language
Boosted Rule-Based Model (C5)	subset: false, no global pruning: false, CF: 0.25, trials: 100	Package ‘C5.0’ R language
Cascade-Correlation Neural Network (CNN)	kernel: sigmoid and gaussian, neuron: 0-10³, candid. 10², epoch 10³ prune to optimal size	Algorithm ‘CNN’ C language
k-Nearest Neighbors (KNN)	model: knn kernel, weighting: gaussian kernel, type: probability	Pack ‘CORElearn’ R language
Probabilistic Neural Network (PNN)	sigma: each var., steps 20, kernel: gaussian, prior prob.: frequency distribution	Algorithm PNN C language
Multilayer Perceptron Network (MLP)	layers: 3, overfitting control: min. holdout val. error over 10% train, function: logistic	Algorithm MLP C language
Random Dec. Forest (Random Forest)	importance: true, proximity: true, number of trees: 300	Pack ‘randomForest’ R language
Single Decision Tree (SDT)	minimum node to split: 3, maximum tree levels: 300, prune to min cross-val error	Algorithm SDT C language
Stochastic Gradient Boosting (TreeBoost)	trees: 300, depth: 8, min size node to split: 10, prune series to min. err, minimum trees in series: 10	Algor. ‘TreeBoost’ C language
Support Vector Machine (SVM)	type: bound-constraint, kernel function: gaussian, sigma: 0.1, C: 24	Package ‘kernlab’ R language

Learning algorithms	5-fold Cross-validation (%)		Hold-out validation (%)
Learning algorithms	Training data	Checking data	Testing data
FRBCS	100	93.5	94.0
C5	100	85.3	86.5
CNN	86.2	76.5	78.5
KNN	96.3	89.1	89.7
PNN	100	95.3	96.1
MLP	94.9	88.5	90.8
Random Forest	100	88.9	89.5
SDT	91.6	72.3	72.3
TreeBoost	100	85.7	87.3
SVM	100	95.9	96.2

Sociedade Brasileira de Matemática Aplicada e Computacional Rua Maestro João Seppe, nº. 900, 16º. andar - Sala 163 , 13561-120 São Carlos - SP, Tel. / Fax: (55 16) 3412-9752 - São Carlos - SP - Brazil
E-mail: sbmac@sbmac.org.br

Acompanhe os números deste periódico no seu leitor de RSS

[1] *Corresponding author: Adriano Bressane - E-mail: adriano.bressane@facens.br.