Ensemble feature selection approach based on feature ranking for rice seed images classification

DIGITAL IMAGE PROCESSING AND COMPUTER GRAPHICS VOLUME: 18 |NUMBER: 3 |2020 |SEPTEMBER Ensemble Feature Selection Approach Based on Feature Ranking for Rice Seed Images Classification Dzi Lam Tran TUAN 1, Thongchai SURINWARANGKOON 2, Kittikhun MEETHONGJAN 3, Vinh Truong HOANG1 1Department of Image Processing and Computer Graphics, Faculty of Computer Science, Ho Chi Minh City Open University, 97 Vo Van Tan Street, 700000 Ho Chi Minh City, Vietnam 2Department of Business Computer, Faculty of Management Science, Suan Sunandha Rajabhat University, 1 U Thong Nok Rd, 10300 Bangkok, Thailand 3Department of Applied Science, Faculty of Science and Technology, Suan Sunandha Rajabhat University, 1 U Thong Nok Rd, 10300 Bangkok, Thailand [email protected], thongc[email protected], kittikh[email protected], [email protected] DOI: 10.15598/aeee.v18i3.3726 Abstract. In smart agriculture, rice variety inspection systems based on computer vision need to be used for recognizing rice seeds instead of using technical experts. In this paper, we have investigated three types of local descriptors, such as Local Binary Pattern (LBP), Histogram of Oriented Gradients (HOG) and GIST to characterize rice seed images. However, this approach raises the curse of dimensionality phenomenon and needs to select the relevant features for a compact and better representation model. A new ensemble feature selection is proposed to represent all useful information collected from different single feature selection methods. The experimental results have shown the efficiency of our proposed method in terms of accuracy. Keywords Ensemble feature selection, feature ranking, feature selection, GIST, HOG, LBP, rice seed image. 1. Introduction Rice is the most important food source of people in many countries including Asia, Africa, Latin America, and the Middle East. Products made from rice, including rice products and indirect products, are indispensable in the daily meals of billions of people around the world. Nowadays, more rice varieties are created with diversified quality and productivity. Different varieties of rice can be mixed during cultivation and trading. We practically need to develop a system to automatically identify rice seeds based on machine vision. Various works have been proposed for automatic inspection and quality control in agriculture [10]. In the past decade, a great number of local image descriptors [13] have been proposed for characterizing images. Each kind of attribute represents the data in a specific space and has precise spatial meaning and statistical properties. Different local descriptors are extracted to create a multi-view image representation, like Local Binary Pattern (LBP), Histogram of Oriented Gradients (HOG), and GIST. Nhat and Hoang [20] present a method to fuse the features extracted from three descriptors (Local Binary Pattern, Histogram of Oriented Gradient and GIST) for facial images classification. The concatenated features are then applied by canonical correlation analysis to have a compact representation before feeding into the classifier. Van and Hoang [24] propose to reduce noisy and irrelevant Local Ternary Pattern (LTP) features and HOG coding on different color spaces for face analysis. Hoai et al. [12] introduce a comparative study of hand-crafted descriptors and Convolutional Neural Networks (CNN) for rice seed images classification. Mebatsion et al. [17] fuse Fourier descriptors and three geometrical features for cereal grains recognition. Duong and Hoang [9] apply to extract rice seed images based on features coded in multiple color spaces using HOG descriptor. Multi-view c 2020 ADVANCES IN ELECTRICAL AND ELECTRONIC ENGINEERING 198 DIGITAL IMAGE PROCESSING AND COMPUTER GRAPHICS VOLUME: 18 |NUMBER: 3 |2020 |SEPTEMBER learning was introduced to complement information between different views. While concatenating different feature sets, it is evident that all the features do not give the same contribution for the learning task and some features might decrease the performance. Thus, feature selection methods are applied as a preprocessing stage to high-dimensional feature space. It involves selecting pertinent and useful features, while avoiding and ignoring redundant and irrelevant information [26]. A novel teacher-student feature selection approach [19] is proposed to find the best representation of data in low dimension. Recently, ensemble feature selection has emerged as a new approach that promises to enhance the robustness and performance. It is the process of performing different feature selection in order to find an optimum subset of features. Instead of using a single selection approach, an ensemble method combines the results of different approaches into a final single subset of features. Seijo-pardo et al. [23] propose to combine different feature selection approaches on heterogeneous data based on a predefined threshold value. Chiew et al. [5] introduce a hybrid ensemble feature selection based on Cumulative Distribution Function gradient. This method can determine an estimation of feature cut-off automatically. Drotar et al. [8] propose a new ensemble feature selection approach methods based on different voting techniques such as plurality, and Borda count. A complete and detailed review of ensemble feature selection methods is introduced in [3]. In this paper, we propose a new ensemble feature selection approach based on multi-view descriptors (LBP, HOG and GIST) extracted from rice seed images. Several feature selection approaches are further investigated and combined to find an optimum subset of features with the purpose to enhance the classification performance. This paper is organized and structured as follows. Section 2. , introduces the feature extracting methods based on three local image descriptors. Section 3. presents a proposed ensemble feature selection framework. Section 4. shows experimental results. Finally, the conclusion is then provided in Sec. 5. . 2. The Feature Extracting Methods This section briefly reviews three local image descriptors used in experiments for feature extraction. 2.1. Local Binary Pattern The LBPP,R(xc, yc)code of each pixel (xc, yc)is calculated by comparing the gray value gcof the central pixel with the gray values {gi}P−1 i=0 of its Pneighbors, as follows [21]: LBPP,R = P−1 X p=0 ω(gp−gc)2p,(1) where gcis the gray value of central, gpis the gray value of P,Ris the radius of the circle, and ω(gp−gc) is defined as: ω(gp−gc) = 1if (gp−gc)≥0, 0otherwise.(2) 2.2. GIST GIST is firstly proposed by Oliva and Torralba [22] in order to classify objects which represent the shape of the object. The primary idea of this approach is based on the Gabor filter: h(x, y) = e −1 2x2 δ2 x +y2 δ2 ye−j2π(u0x+v0y).(3) For each (δx,δy) of the image via the Gabor filter, we obtain all the image elements that are close to the point color (u0x+v0y). The result of the calculated Vector GIST will have many dimensions. To reduce the size of the vector, we averaged each 4×4 grid of the above results. Each image also configures a Gabor filter with 4 scales and 8 directions (orientations), creating 32 characteristic maps of the same size. 2.3. Histograms of Oriented Gradient HOG descriptors are applied for different tasks in machine vision [7] such as human detection [6]. HOG feature is extracted by counting the occurrences of gradient orientation based on the gradient angle and the gradient magnitude of local patches of an image. The gradient angle and magnitude at each pixel is computed in an 8×8pixels patch. Next, 64 gradient feature vectors are divided into 9 angular bins 0–180◦(20◦each). The gradient magnitude Tand angle Kat each position (k,h) from an image Jare computed as follows: ∆k=|J(k−1, h)−J(k+ 1, h)|.(4) ∆h=|J(k, h −1) −J(k, h + 1)|.(5) T(k, h) = q∆2 i+ ∆2 j.(6) K(k, h) = tan−1∆k ∆j.(7) c 2020 ADVANCES IN ELECTRICAL AND ELECTRONIC ENGINEERING 199 DIGITAL IMAGE PROCESSING AND COMPUTER GRAPHICS VOLUME: 18 |NUMBER: 3 |2020 |SEPTEMBER Data (featured description) Training set Testing set Feature Selections Final model Acc Top 2 split top Ranking Acc Top 3 split top Ranking Acc Top 1 combine New ranking split top Ranking combine Feature Selection Final model Acc Rankings last Ranking Fig. 1: The proposed ensemble feature selection approach. 3. Ensemble Feature Selection The dimension reduction has several advantages and impacts on data storage, generalization capability and computing time. Based on the availability of supervised information (i.e, class labels), feature selection techniques can be grouped into two large categories: supervised and unsupervised context [1]. Additionally, different strategies of feature selection are proposed based on evaluation processes such as filter, wrapper and hybrid methods [11]. Hybrid approaches incorporate both filter and wrapper into a single structure, in order to give an effective solution for dimensionality reduction [4]. In order to study the contribution of feature selection approaches for rice seed images classification, we propose to apply several selection approaches based on images represented by multi-view descriptors. In the following subsection, we will shortly present the common feature selection methods applied in supervised learning context. •LASSO (Least Absolute Shrinkage and Selection Operator) allows to compute feature selection based on the assumption of linear dependency between input features and output values. Lasso minimizes the sum of squares of residuals when the sum of the absolute values of the regression coefficients is less than a constant, which yields certain strict regression coefficients equal to 0 [4] and [25]. •mRMR (Maximum Relevance and Minimum Redundancy) is a mutual information based feature selection criterion, or distance/similarity scores to select features. The aim is to penalize a feature’s relevance by its redundancy in the presence of the other selected features [27]. •ReliefF [15] is extended from Relief [14] to support multiclass problems. ReliefF seems to be a promising heuristic function that may overcome the myopia of current inductive learning algorithms. Kira and Rendell used ReliefF as a preprocessor to eliminate irrelevant attributes from data description before learning. ReliefF is general, relatively efficient, and reliable enough to guide the search in the learning process [16]. •CFS (Correlation Feature Selection) mainly applies heuristic methods to evaluate the effect of a single feature corresponding to each group in order to obtain the optimal subset of attributes. c 2020 ADVANCES IN ELECTRICAL AND ELECTRONIC ENGINEERING 200 DIGITAL IMAGE PROCESSING AND COMPUTER GRAPHICS VOLUME: 18 |NUMBER: 3 |2020 |SEPTEMBER •Fisher [2] identifies a subset of features so that the distances between samples in different classes are as large as possible, while the distances between samples in the same class are as small as possible. Fisher selects the top ranked features according to its scores. •ILFS (Infinite Latent Feature Selection) is a technique consists of three steps such as preprocessing, feature weighting based on a fully connected graph in each node that connect all features. Finally, energy scores of the path length are calculated, then rank its correspondence with the feature [18]. Figure 1 present the proposed ensemble feature selection framework. Each individual feature selection approach has its pros and cons, the aim of this proposition is to combine the pros of different methods to boost the performance in terms of accuracy. We propose to apply three independent feature selection methods to select the "best" subset of features. Then, a new ranking method is applied for combined feature space. This can increase the dimension space but it allows to collect relevant features determined by different selection methods. The meaning behind is to select the most relevant features so that we have to apply a final ranking to eliminate the redundant and noisy features. 4. Experimental Results 4.1. Experimental Setup (a) BC-15. (b) Nep-87. (c) Huong thom 1. (d) Thien uu 8. (e) Q-5. (f) Xi-23. Fig. 2: The proposed of ensemble feature selection. The rice seed images database comprises six rice seed varieties in the northern Vietnam (illustrated in Fig. 2) [9]. We apply the 1-NN and SVM classifiers to evaluate the classification performance via accuracy rate. A half of the database is selected for the training set and the rest is used for the testing set. We use Hold-out method with ratio (1/2 and 1/2) and split the training and testing set by chessboard decomposition. All experiments are implemented and simulated by Matlab 2019a and conducted on a PC with a configuration of a CPU Xeon 3.08 GHz, 64 GBs of RAM. 4.2. Results Table 1 shows the accuracy obtained by 1-NN and SVM classifier when no feature selection approach is applied. The first column indicates the features used for representing images. We use three individual local descriptors namely LBP, GIST, and HOG and the concatenation of "LBP + GIST" features. The second column indicates the number of features (or dimension) corresponding to features type. The third and fourth columns show the accuracy obtained by 1-NN and SVM classifier. We observe that the multi-view by concatenating multiple features gives better performance, however it increases the dimension. Hence, the performance of SVM classifier is better than 1-NN classifier with 94.7 % of accuracy. Tab. 1: Classification performance without selection approach for different types of features. Features Dimension 1-NN SVM LBP 768 53.0 77.0 GIST 512 69.4 88.3 HOG 21.384 71.5 94.7 LBP + GIST 1.280 70.5 91.7 The following tables and figures illustrate in detailed the classification in single or multi-view based on three descriptors: •LBP: Table 2, Fig. 3(a) and Fig. 3(b), •GIST: Table 4, Fig. 4(a) and Fig. 4(b), •HOG: Table 5, Fig. 5(a) and Fig. 5(b), •LBP + GIST: Table 3, Fig. 6(a) and Fig. 6(b). Table 2 and Fig. 3 show that the classification performance reach 53.0 % by 1-NN classifier on LBP descriptor. After using 6 different feature selection approaches, we obtain three best candidates with descendant accuracy such as mRMR (59.0 %), ILFS (58.4 %) and ReliefF (54.2 %). Based on the proposed method illustrated in Fig. 1, the 85 % percentage of selected features by ReliefF is combined with 43 % of selected feature determined by ILFS method. We obtain the new subset of features which is calculated as follows: (768 ·0.85) + (768 ·0.43) = 983 dim.(8) c 2020 ADVANCES IN ELECTRICAL AND ELECTRONIC ENGINEERING 201 DIGITAL IMAGE PROCESSING AND COMPUTER GRAPHICS VOLUME: 18 |NUMBER: 3 |2020 |SEPTEMBER 0 10 20 30 40 50 60 70 80 90 100 Number of selected features 20 25 30 35 40 45 50 55 60 Accuracy 1-NN classifier on LBP features CFS FISHER ILFS LASSO MRMR RELIEFF (a) 1-NN. 0 10 20 30 40 50 60 70 80 90 100 Number of selected features 30 40 50 60 70 80 90 Accuracy SVM classifier on LBP features CFS FISHER ILFS LASSO MRMR RELIEFF (b) SVM. Fig. 3: 1-NN (a) and SVM (b) classifier on LBP features. 0 10 20 30 40 50 60 70 80 90 100 Number of selected features 25 30 35 40 45 50 55 60 65 70 75 Accuracy 1-NN classifier on GIST features CFS FISHER ILFS LASSO MRMR RELIEFF (a) 1-NN. 0 10 20 30 40 50 60 70 80 90 100 Number of selected features 30 40 50 60 70 80 90 100 Accuracy SVM classifier on GIST features CFS FISHER ILFS LASSO MRMR RELIEFF (b) SVM. Fig. 4: 1-NN (a) and SVM (b) classifier on GIST features. 0 10 20 30 40 50 60 70 80 90 100 Number of selected features 45 50 55 60 65 70 75 80 Accuracy 1-NN classifier on HOG features CFS FISHER ILFS LASSO MRMR RELIEFF (a) 1-NN. 0 10 20 30 40 50 60 70 80 90 100 Number of selected features 65 70 75 80 85 90 95 100 Accuracy SVM classifier on HOG features CFS FISHER ILFS LASSO MRMR RELIEFF (b) SVM. Fig. 5: 1-NN (a) and SVM (b) classifier on HOG features. c 2020 ADVANCES IN ELECTRICAL AND ELECTRONIC ENGINEERING 202 DIGITAL IMAGE PROCESSING AND COMPUTER GRAPHICS VOLUME: 18 |NUMBER: 3 |2020 |SEPTEMBER 0 10 20 30 40 50 60 70 80 90 100 Number of selected features 25 30 35 40 45 50 55 60 65 70 75 Accuracy 1-NN classifier on LBP + GIST features CFS FISHER ILFS LASSO MRMR RELIEFF (a) 1-NN. 0 10 20 30 40 50 60 70 80 90 100 Number of selected features 30 40 50 60 70 80 90 100 Accuracy SVM classifier on LBP + GIST features CFS FISHER ILFS LASSO MRMR RELIEFF (b) SVM. Fig. 6: 1-NN (a) and SVM (b) classifier on LBP + GIST features. Tab. 2: LBP features - classification performance based on different feature selection methods with 1-NN and SVM classifier. ACC: accuracy, Dim: dimension, id %: percentage of selected features, ⩾id %: percentage of selected features with accuracy equal to all features are used. LBP Dim 1-NN SVM ACC Max ACC ACC Max ACC 100 % ⩾id % Dim Max id % Dim 100 % ⩾id % Dim Max id % Dim Fisher 768 53.0 80 614 53.6 96 737 77.0 84 645 77.4 87 668 mRMR 768 53.0 11 84 59.0 28 215 77.0 22 169 81.8 37 284 ReliefF 768 53.0 74 568 54.2 85 653 77.0 97 745 77.0 97 745 Ilfs 768 53.0 12 92 58.4 43 330 77.0 19 146 81.6 40 307 Cfs 768 53.0 90 691 52.3 96 737 77.0 96 737 77.1 96 737 Lasso 768 53.0 94 722 53.1 94 722 77.0 100 768 77.0 100 768 Tab. 3: LBP + GIST features - classification performance based on different feature selection methods with 1-NN and SVM classifier. ACC: accuracy, Dim: dimension, id %: percentage of selected features, ⩾id %: percentage of selected features with accuracy equal to all features are used. LBP Dim 1-NN SVM + ACC Max ACC ACC Max ACC GIST 100 % ⩾id % Dim Max id % Dim 100 % ⩾id % Dim Max id % Dim Fisher 1280 70.5 88 1126 70.7 88 1126 91.7 100 1280 91.7 100 1280 mRMR 1280 70.5 31 397 72.7 52 666 91.7 40 512 92.4 69 883 ReliefF 1280 70.5 49 627 73.8 68 870 91.7 94 1203 91.9 96 1229 Ilfs 1280 70.5 27 346 72.4 72 922 91.7 41 525 94.2 58 742 Cfs 1280 70.5 59 755 70.9 94 1203 91.7 98 1254 91.7 98 1254 Lasso 1280 70.5 10 128 70.9 10 128 91.7 98 1254 91.7 98 1254 Tab. 4: LGIST features - classification performance based on different feature selection method with 1-NN and SVM classifier. ACC: accuracy, Dim: dimension, id %: percentage of selected features, ⩾id %: percentage of selected features with accuracy equal to all features are used. Dim GIST 1-NN SVM ACC Max ACC ACC Max ACC 100 % ⩾id % Dim Max id % Dim 100 % ⩾id % Dim Max id % Dim Fisher 512 69.4 42 215 70.2 47 241 88.3 98 502 88.3 98 502 mRMR 512 69.4 39 200 71.4 53 271 88.3 48 246 90.8 66 338 ReliefF 512 69.4 21 108 73.4 70 358 88.3 36 184 90.2 46 236 Ilfs 512 69.4 49 251 70.0 79 404 88.3 99 507 88.4 99 507 Cfs 512 69.4 38 195 71.2 75 384 88.3 49 251 90.2 82 420 Lasso 512 69.4 40 205 69.7 99 507 88.3 58 297 90.6 78 399 c 2020 ADVANCES IN ELECTRICAL AND ELECTRONIC ENGINEERING 203 DIGITAL IMAGE PROCESSING AND COMPUTER GRAPHICS VOLUME: 18 |NUMBER: 3 |2020 |SEPTEMBER Tab. 5: HOG features - classification performance based on different feature selection method with 1-NN and SVM classifier. ACC: accuracy, Dim: dimension, id %: percentage of selected features, ⩾id %: percentage of selected features with accuracy equal to all features are used. HOG Dim 1-NN SVM ACC Max ACC ACC Max ACC 100 % ⩾id % Dim Max id % Dim 100 % ⩾id % Dim Max id % Dim Fisher 21384 71.5 20 4277 73.2 27 5774 94.8 85 18176 94.8 99 21170 mRMR 21384 71.5 8 1711 73.9 14 2994 94.8 100 21384 94.8 100 21384 ReliefF 21384 71.5 2 428 74.4 3642 94.8 100 21384 94.8 100 21384 Ilfs 21384 71.5 100 21384 71.5 100 21384 94.8 100 21384 94.8 100 21384 Cfs 21384 71.5 8 1711 72.9 21 4491 94.8 51 10906 95.1 74 15824 Lasso 21384 71.5 9 1925 75.5 19 4063 94.8 100 21384 94.8 100 21384 Tab. 6: The classification results obtained by single and ensemble feature selection. Classifier Dataset Single FS Multi FS Description Dim ACC ACC max Acc Dim Pair Dim Ranker full without FS of FSs (%) full (%) (%) 1-NN LBP 768 53.0 59.0 60.0 432 Ilfs, 983 mRMR ReliefF GIST 512 69.4 73.0 74.6 261 mRMR 655 ReliefF Cfs HOG 21384 71.5 75.5 79.3 3416 mRMR 3635 ReliefF ReliefF LBP + GIST 1280 70.5 73.8 77.1 698 mRMR 1587 ReliefF Ilfs SVM LBP 768 77.0 81.8 82.4 544 Ilfs 591 mRMR mRMR GIST 512 88.3 90.8 91.4 1076 Ilfs 1346 mRMRmRMR Fisher LBP + GIST 1280 91.7 94.2 94.0 1246 mRMR 2112 Ilfs ReliefF So, we combine two best subset of features determined by ReliefF and ILFS with a feature space equal to 983. Next, this vector is applied again by mRMR method and 1-NN classifier to remove irrelevant features. Table 6 presents the comparison of a single and ensemble feature selection framework. We observe that the ensemble method outperforms single feature selection method for all kinds of features with 1-NN classifier. For example, we increase 1 % of accuracy compared to a single feature selection method and increase 7 % compared with the classification when no selection method is applied. Similar experimental results are obtained by using SVM classifier on single view descriptor. In terms of dimension, we increase the feature space by combining and selecting useful information of different single feature selection methods. Compared with the aims based on accuracy or time computing, an appropriate approach for such demand has to be chosen. 5. Conclusion In this paper, we introduced a new ensemble feature selection approach by combining multiple single feature selection methods. A pre-selected subset of features is first determined by considering feature selection and associated classifier. Multiple subsets are then combined to form a final feature space and then applied feature selection method again to eliminate noisy and redundant features. The experimental results on the VNRICE dataset for rice seed images classification have shown the efficiency of the proposed approach. The future of this work is to determine an appropriate selection method based on each attribute and using different strategies to combine the final feature vector resulting from a single feature selection method. Acknowledgment This work was supported by Suan Sunandha Rajabhat University, Thailand References [1] BENABDESLEM, K. and M. HINDAWI. Constrained Laplacian Score for Semi-supervised Feature Selection. In: Joint European Conference on Machine Learning and Knowledge Discovery in Databases. Berlin: Springer, 2011, pp. 204–218. c 2020 ADVANCES IN ELECTRICAL AND ELECTRONIC ENGINEERING 204 DIGITAL IMAGE PROCESSING AND COMPUTER GRAPHICS VOLUME: 18 |NUMBER: 3 |2020 |SEPTEMBER ISBN 978-3-642-23780-5. DOI: 10.1007/978-3-64223780-5_23. [2] BISHOP, C. M. Neural Networks for Pattern Recognition. 1st ed. Oxford: Oxford university press, 1996. ISBN 978-0-198-53864-6. [3] BOLON-CANEDO, V. and A. ALONSOBETANZOS. Ensembles for feature selection: A review and future trends. Information Fusion. 2019, vol. 52, iss. 1, pp. 1–12. ISSN 1566-2535. DOI: 10.1016/j.inffus.2018.11.008. [4] CAI, J., J. LUO, S. WANG and S. YANG. Feature selection in machine learning: A new perspective. Neurocomputing. 2018, vol. 300, iss. 1, pp. 70–79. ISSN 0925-2312. DOI: 10.1016/j.neucom.2017.11.077. [5] CHIEW, K. L., C. L. TAN, K. WONG, K. S. C. YONG and W. K. TIONG. A new hybrid ensemble feature selection framework for machine learning-based phishing detection system. Information Sciences. 2019, vol. 484, iss. 1, pp. 153–166. ISSN 0020-0255. DOI: 10.1016/j.ins.2019.01.064. [6] DALAL, N. and B. TRIGGS. Histograms of oriented gradients for human detection. In: 2005 IEEE computer society conference on computer vision and pattern recognition (CVPR’05). San Diego: IEEE, 2005, pp. 886–893. IEEE. ISBN 07695-2372-2. DOI: 10.1109/CVPR.2005.177. [7] DENIZ, O., G. BUENO, J. SALIDO and F. DE LA TORRE. Face recognition using Histograms of Oriented Gradients. Pattern Recognition Letters. 2011, vol. 32, iss. 12, pp. 1598–1603. ISSN 1598-1603. DOI: 10.1016/j.patrec.2011.01.004. [8] DROTAR, P., M. GAZDA and L. VOKOROKOS. Ensemble feature selection using election methods and ranker clustering. Information Sciences. 2019, vol. 480, iss. 1, pp. 365–380. ISSN 0020-0255. DOI: 10.1016/j.ins.2018.12.033. [9] DUONG, H.-T. and V. T. HOANG. Dimensionality Reduction Based on Feature Selection for Rice Varieties Recognition. In: 2019 4th International Conference on Information Technology (InCIT). Bangkok: IEEE, 2019, pp. 199–202. ISBN 978-17281-1019-6. DOI: 10.1109/INCIT.2019.8912121. [10] GOMES, J. F. S. and F. R. LETA. Applications of computer vision techniques in the agriculture and food industry: a review. European Food Research and Technology. 2012, vol. 235, iss. 6, pp. 989–1000. ISSN 1438-2385. DOI: 10.1007/s00217-012-1844-2. [11] GUYON, I. and A. ELISSEEFF. An Introduction to Variable and Feature Selection. Journal of Machine Learning Research. 2003, vol. 3, iss. 7, pp. 1157–1182. ISSN 1533-7928. DOI: 10.5555/944919.944968. [12] D. P. VAN HOAI, T. SURINWARANGKOON, V. T. HOANG, H.-T. DUONG and K. MEETHONGJAN. A Comparative Study of Rice Variety Classification based on Deep Learning and Hand-crafted Features. ECTI Transactions on Computer and Information Technology (ECTI-CIT). 2020, vol. 14, iss. 1, pp. 1–10. ISSN 2286-9131. DOI: 10.37936/ecticit.2020141.204170. [13] HUMEAU-HEURTIER, A. Texture Feature Extraction Methods: A Survey. IEEE Access. 2019, vol. 7, iss. 1, pp. 8975–9000. ISSN 2169-3536. DOI: 10.1109/ACCESS.2018.2890743. [14] KIRA, K. and L. A. RENDELL. A Practical Approach to Feature Selection. In: Machine Learning Proceedings 1992. Aberdeen: Elsevier, 1992, pp. 249–256. ISBN 978-1-558-60247-2. DOI: 10.1016/B978-1-55860-247-2.50037-1. [15] KONONENKO, I. Estimating attributes: Analysis and extensions of RELIEF. In: European Conference on Machine Learning. Springer, 1994, pp. 171–182. ISBN 978-3-540-48365-6. DOI: 10.1007/3-540-57868-4_57. [16] KONONENKO, I., E. SIMEC and M. ROBNIKSIKONJA. Overcoming the Myopia of Inductive Learning Algorithms with RELIEFF. Applied Intelligence. 1997, vol. 7, iss. 1, pp. 39–55. ISSN 1573-7497. DOI: 10.1023/A:1008280620621. [17] MEBATSION, H. K., J. PALIWAL and D. S. JAYAS. Automatic classification of nontouching cereal grains in digital images using limited morphological and color features. Computers and Electronics in Agriculture. 2013, vol. 90, iss. 1, pp. 99–105. ISSN 0168-1699. DOI: 10.1016/j.compag.2012.09.007. [18] MIFTAHUSHUDUR, T., C. B. A. WAEL and T. PRALUDI. Infinite Latent Feature Selection Technique for Hyperspectral Image Classification. Jurnal Elektronika dan Telekomunikasi. 2019, vol. 19, iss. 1, pp. 32–37. ISSN 2527-9955. DOI: 10.14203/jet.v19.32-37. [19] MIRZAEI, A., V. POURAHMADI, M. SOLTANI and H. SHEIKHZADEH. Deep feature selection using a teacher-student network. Neurocomputing. 2020, vol. 383, iss. 1, pp. 396–408. ISSN 0925-2312. DOI: 10.1016/j.neucom.2019.12.017. c 2020 ADVANCES IN ELECTRICAL AND ELECTRONIC ENGINEERING 205 DIGITAL IMAGE PROCESSING AND COMPUTER GRAPHICS VOLUME: 18 |NUMBER: 3 |2020 |SEPTEMBER [20] NHAT, H. T. M. and V. T. HOANG. Feature fusion by using LBP, HOG, GIST descriptors and Canonical Correlation Analysis for face recognition. In: 2019 26th International Conference on Telecommunications (ICT). Hanoi: IEEE, 2019, pp. 371–375. ISBN 978-1-7281-02733. DOI: 10.1109/ICT.2019.8798816. [21] OJALA, T., M. PIETIKAINEN and T. MAENPAA. A Generalized Local Binary Pattern Operator for Multiresolution Gray Scale and Rotation Invariant Texture Classification. In: International Conference on Advances in Pattern Recognition. Rio de Janeiro: Springer, 2001, pp. 399– 408. ISBN 978-3-540-41767-5. DOI: 10.1007/3540-44732-6_41. [22] OLIVA, A. and A. TORRALBA. Modeling the Shape of the Scene: A Holistic Representation of the Spatial Envelope. International Journal of Computer Vision. 2001, vol. 42, iss. 3, pp. 145–175. ISSN 0920-5691. DOI: 10.1023/A:1011139631724. [23] SEIJO-PARDO, B., I. PORTO-DIAZ, V. BOLON-CANEDO and A. ALONSOBETANZOS. Ensemble feature selection: Homogeneous and heterogeneous approaches. Knowledge-Based Systems. 2017, vol. 118, iss. 1, pp. 124–139. ISSN 0950-7051. DOI: 10.1016/j.knosys.2016.11.017. [24] VAN, T. N. and V. T. HOANG. Kinship Verification based on Local Binary Pattern features coding in different color space. In: 2019 26th International Conference on Telecommunications (ICT). Hanoi: IEEE, 2019, pp. 376–380. ISBN 978-17281-0273-3. DOI: 10.1109/ICT.2019.8798781. [25] YAMADA, M., W. JITKRITTUM, L. SIGAL, E. P. XING and M. SUGIYAMA. HighDimensional Feature Selection by Feature-Wise Kernelized Lasso. Neural Computation. 2014, vol. 26, iss. 1, pp. 185–207. ISSN 1530-888X. DOI: 10.1162/NECO_a_00537. [26] ZHANG, R., F. NIE, X. LI and X. WEI. Feature selection with multi-view data: A survey. Information Fusion. 2019, vol. 50, iss. 1, pp. 158–167. ISSN 1566-2535. DOI: 10.1016/j.inffus.2018.11.019. [27] ZHAO, Z., R. ANAND and M. WANG. Maximum Relevance and Minimum Redundancy Feature Selection Methods for a Marketing Machine Learning Platform. In: 2019 IEEE International Conference on Data Science and Advanced Analytics (DSAA). Washington: IEEE, 2019, pp. 442–452. ISBN 978-1-7281-4493-1. DOI: 10.1109/DSAA.2019.00059. About Authors Dzi Lam Tran TUAN is a graduate student of Ho Chi Minh City Open University, Vietnam. He obtained his M.Sc. degree in Computer Science in 2019. His research interests include feature selection and image analysis. Thongchai SURINWARANGKOON was born in Suratthani, Thailand in 1972. He obtained a B.Sc. degree in mathematics from Chiang Mai University, Thailand in 1995. He received M.Sc. degree in management of information technology from Walailak University, Thailand in 2005 and Ph.D. in information technology from King Mongkut’s University of Technology North Bangkok, Thailand in 2014. He is now a Lecturer in the Department of Business Computer, Suan Sunandha Rajabhat University, Bangkok, Thailand. His research interests include digital image processing, machine learning, artificial intelligence, business intelligence, and mobile application development for business. Kittikhun MEETHONGJAN is a full lecturer in the Computer Science Program, Head of Apply Science Department, Faculty of Science and Technology, Suan Sunandha Rajabhat University (SSRU), Thailand. He received his B.Sc. Degree in Computer Science from SSRU in 1990, a M.Sc. degree in computer science from King Mongkut’s University of Technology Thonburi (KMUTT) in 2000 and a Ph.D. degree in Computer Graphic from the University of Technology, Malaysia, in 2013. His research interest is Computer Graphic, Image Processing, Artificial Intelligent, Biometrics, Pattern Recognition, Soft Computing techniques and, application. Currently, he is an advisor of Master and Ph.D. student in Forensic Science of SSRU. Vinh Truong HOANG received his M.Sc. degree from the University of Montpellier in 2009 and his Ph.D. degree in computer science from the University of the Littoral Opal Coast, France. He is currently an assistant professor and Head of Image Processing and Computer Graphics Department at the Ho Chi Minh City Open University, Vietnam. His research interests include image analysis and feature selection. c 2020 ADVANCES IN ELECTRICAL AND ELECTRONIC ENGINEERING 206