# Histogram-Based Descriptor Subset Selection for Visual Recognition of Industrial Parts

^{1}

^{2}

^{*}

## Abstract

**:**

## 1. Introduction

## 2. Background

#### 2.1. Features Descriptors

#### 2.2. Classifiers

#### 2.3. Feature Selection

## 3. Proposed Approach

#### 3.1. Classification

#### 3.2. Feature Selection Techniques

Algorithm 1: Sequential Forward Subset Selection |

Algorithm 2: Sequential Backward Subset Selection |

#### 3.3. Evaluation Measure

#### 3.4. Full Pipeline

## 4. Experiments and Results

#### 4.1. Forward Subset Selection

#### 4.2. Backward Subset Selection

#### 4.3. Comparative between Methods

## 5. Conclusions

## Author Contributions

## Funding

## Conflicts of Interest

## Appendix A. Parameters

Algorithm | Parameter | Description of the Parameter | Value |
---|---|---|---|

KNN | K | number of neighbors | 1 |

SVM SMO | C | parameter C | 1.0 |

L | tolerance | 0.001 | |

P | epsilon for round-off error | $1.0\times {10}^{-12}$ | |

N | Normalization | true | |

V | calibration folds | −1 | |

K | Kernel | PolyKernel | |

C PolyKernel | Cache size of the kernel | 250,007 | |

E PolyKernel | Exponent value of the kernel | 2.0 | |

SVM SGD | M | Multiclass type | 1-against-all |

F | Loss function | hinge loss | |

L | Learning rate | 0.001 | |

R | Regulation constant | 0.0001 | |

E | Number of epochs to perform | 500 | |

C | Epsilon threshold for loss function | 0.001 | |

RC | W | The base classifier to be used | RandomTree |

K | Number of choosen attributes in the RandomTree | $int(lo{g}_{2}\left(predictors\right)+1)$ | |

M RandomTree | Minimum total weight in a leaf | 1.0 | |

V RandomTree | Minimum proportion of variance | 0.001 | |

RF | P | Size of each bag | 100 |

I | Number of iterations | 100 | |

K | Number of randomly choosen attributes | $int(lo{g}_{2}\left(predictors\right)+1)$ | |

M RandomTree | Minimum total weight in a leaf | 1.0 | |

V RandomTree | Minimum proportion of variance | 0.001 | |

Bagging | P | Size of each bag | 100 |

I | Number of iterations | 10 | |

W | The base classifier to be used | REPTree (Fast Decision Tree) | |

M REPTree | Minimum total weight in a leaf | 2 | |

V REPTree | Minimum proportion of variance | 0.001 | |

N REPTree | Amount of data used for prunning | 3 | |

L REPTree | Maximum depth of the tree | −1 (no restriction) | |

I REPTree | Initial class value count | 0.0 |

## References

- Teichmann, M.; Weber, M.; Zoellner, M.; Cipolla, R.; Urtasun, R. MultiNet: Real-time Joint Semantic Reasoning for Autonomous Driving. arXiv
**2016**, arXiv:cs.CV/1612.07695. [Google Scholar] - Ronneberger, O.; Fischer, P.; Brox, T. U-Net: Convolutional Networks for Biomedical Image Segmentation. arXiv
**2015**, arXiv:cs.CV/1505.04597. [Google Scholar] - Yan, M.; Zhao, M.; Xu, Z.; Zhang, Q.; Wang, G.; Su, Z. VarGFaceNet: An Efficient Variable Group Convolutional Neural Network for Lightweight Face Recognition. arXiv
**2019**, arXiv:cs.CV/1910.04985. [Google Scholar] - Liu, Y.; Wang, Y.; Wang, S.; Liang, T.; Zhao, Q.; Tang, Z.; Ling, H. CBNet: A Novel Composite Backbone Network Architecture for Object Detection. arXiv
**2019**, arXiv:cs.CV/1909.03625. [Google Scholar] - Yuan, Y.; Chen, X.; Wang, J. Object-Contextual Representations for Semantic Segmentation. arXiv
**2019**, arXiv:cs.CV/1909.11065. [Google Scholar] - Gómez, A.; de la Fuente, D.C.; García, N.; Rosillo, R.; Puche, J. A vision of industry 4.0 from an artificial intelligence point of view. In Proceedings of the 18th International Conference on Artificial Intelligence, Varna, Bulgaria, 25–28 July 2016; p. 407. [Google Scholar]
- Chen, C.; Liu, M.Y.; Tuzel, O.; Xiao, J. R-CNN for Small Object Detection. In Computer Vision—ACCV 2016; Lai, S.H., Lepetit, V., Nishino, K., Sato, Y., Eds.; Springer International Publishing: Cham, Switzerland, 2017; pp. 214–230. [Google Scholar]
- Xie, Q.; Luong, M.T.; Hovy, E.; Le, Q.V. Self-training with Noisy Student improves ImageNet classification. arXiv
**2020**, arXiv:1911.04252. [Google Scholar] - Cubuk, E.D.; Zoph, B.; Mane, D.; Vasudevan, V.; Le, Q.V. AutoAugment: Learning Augmentation Policies from Data. arXiv
**2019**, arXiv:1805.09501. [Google Scholar] - Kolesnikov, A.; Beyer, L.; Zhai, X.; Puigcerver, J.; Yung, J.; Gelly, S.; Houlsby, N. Large Scale Learning of General Visual Representations for Transfer. arXiv
**2019**, arXiv:1912.11370. [Google Scholar] - Wang, J.; Chen, Y.; Yu, H.; Huang, M.; Yang, Q. Easy Transfer Learning By Exploiting Intra-domain Structures. arXiv
**2019**, arXiv:cs.LG/1904.01376. [Google Scholar] - Nannia, L.; Ghidonia, S.; Brahnamb, S.; Le, Q.V. Handcrafted vs. non-handcrafted features for computer vision classification. Pattern Recognit.
**2017**, 71, 158–172. [Google Scholar] - Lowe, D.G. Object recognition from local scale-invariant features. In Proceedings of the Seventh IEEE International Conference on Computer Vision, Kerkyra, Greece, 20–27 September 1999; Volume 2, pp. 1150–1157. [Google Scholar]
- Bay, H.; Tuytelaars, T.; Van Gool, L. SURF: Speeded Up Robust Features. In Proceedings of the European Conference on Computer Vision, Graz, Austria, 7–13 May 2006; pp. 404–417. [Google Scholar]
- Ojala, T.; Pietikäinen, M.; Harwood, D. A comparative study of texture measures with classification based on featured distributions. Pattern Recognit.
**1996**, 29, 51–59. [Google Scholar] [CrossRef] - Fernández, A.; Álvarez, M.X.; Bianconi, F. Texture Description Through Histograms of Equivalent Patterns. J. Math. Imaging Vis.
**2013**, 45, 76–102. [Google Scholar] [CrossRef][Green Version] - Merino, I.; Azpiazu, J.; Remazeilles, A.; Sierra, B. 2D Features-based Detector and Descriptor Selection System for Hierarchical Recognition of Industrial Parts. IJAIA
**2019**, 10, 1–13. [Google Scholar] [CrossRef] - Haralick, R.M.; Shanmugam, K.; Dinstein, I. Textural Features for Image Classification. IEEE Trans. Syst. Man Cybern.
**1973**, SMC-3, 610–621. [Google Scholar] [CrossRef][Green Version] - Wang, L.; He, D.C. Texture classification using texture spectrum. Pattern Recognit.
**1990**, 23, 905–910. [Google Scholar] - Madrid-Cuevas, F.J.; Medina, R.; Prieto, M.; Fernández, N.L.; Carmona, A. Simplified Texture Unit: A New Descriptor of the Local Texture in Gray-Level Images. In Pattern Recognition and Image Analysis; Springer: Berlin/Heidelberg, Germany, 2003; Volume 2652, pp. 470–477. [Google Scholar]
- Xu, B.; Gong, P.; Seto, E.; Spear, R. Comparison of Gray-Level Reduction and Different Texture Spectrum Encoding Methods for Land-Use Classification Using a Panchromatic Ikonos Image. Photogramm. Eng. Remote Sens.
**2003**, 69, 529–536. [Google Scholar] - Zhang, W.; Shan, S.; Gao, W.; Chen, X.; Zhang, H. Local Gabor binary pattern histogram sequence (LGBPHS): A novel non-statistical model for face representation and recognition. In Proceedings of the Tenth IEEE International Conference on Computer Vision (ICCV’05) Volume 1, Beijing, China, 17–21 October 2005; Volume 1, pp. 786–791. [Google Scholar]
- Tan, X.; Triggs, W. Enhanced local texture feature sets for face recognition under difficult lighting conditions. IEEE Trans. Image Process.
**2010**, 19, 1635–1650. [Google Scholar] - Fernández, A.; Álvarez, M.X.; Bianconi, F. Image classification with binary gradient contours. Opt. Lasers Eng.
**2011**, 49, 1177–1184. [Google Scholar] - Hussain, S.U.; Napoléon, T.; Jurie, F. Face Recognition using Local Quantized Patterns. In Procedings of the British Machine Vision Conference 2012; British Machine Vision Association: Guildford, UK, 2012; pp. 99.1–99.11. [Google Scholar]
- Chen, J.; Shan, S.; He, C.; Zhao, G.; Pietikainen, M.; Chen, X.; Gao, W. WLD: A Robust Local Image Descriptor. IEEE Trans. Pattern Anal. Mach. Intell.
**2010**, 32, 1705–1720. [Google Scholar] [CrossRef] - Dalal, N.; Triggs, B. Histograms of oriented gradients for human detection. In Proceedings of the 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR’05), San Diego, CA, USA, 20–25 June 2005; IEEE: Piscataway, NJ, USA, 2005; Volume 1, pp. 886–893. [Google Scholar]
- Cover, T.; Hart, P. Nearest neighbor pattern classification. IEEE Trans. Inf. Theory
**1967**, 13, 21–27. [Google Scholar] - Rish, I. An empirical study of the naive Bayes classifier. In IJCAI 2001 Workshop on Empirical Methods in Artificial Intelligence; IBM: New York, NY, USA, 2001; Volume 3, pp. 41–46. [Google Scholar]
- Breiman, L. Random forests. Mach. Learn.
**2001**, 45, 5–32. [Google Scholar] [CrossRef][Green Version] - Cortes, C.; Vapnik, V. Support-vector networks. Mach. Learn.
**1995**, 20, 273–297. [Google Scholar] [CrossRef] - Lira, M.M.S.; de Aquino, R.R.B.; Ferreira, A.A.; Carvalho, M.A.; Neto, O.N.; Santos, G.S.M. Combining Multiple Artificial Neural Networks Using Random Committee to Decide upon Electrical Disturbance Classification. In Proceedings of the 2007 International Joint Conference on Neural Networks, Orlando, FL, USA, 12–17 August 2007; pp. 2863–2868. [Google Scholar]
- Breiman, L. Bagging predictors. Mach. Learn.
**1996**, 24, 123–140. [Google Scholar] [CrossRef][Green Version] - Shahlaei, M. Descriptor Selection Methods in Quantitative Structure–Activity Relationship Studies: A Review Study. Chem. Rev.
**2013**, 113, 8093–8103. [Google Scholar] [CrossRef] [PubMed] - Rasines, I.; Remazeilles, A.; Bengoa, P.M.I. Feature selection for hand pose recognition in human-robot object exchange scenario. In Proceedings of the 2014 IEEE Emerging Technology and Factory Automation (ETFA), Barcelona, Spain, 16–19 September 2014; pp. 1–8. [Google Scholar]
- Molina, L.; Belanche, L.; Nebot, A. Feature selection algorithms: A survey and experimental evaluation. In Proceedings of the 2002 IEEE International Conference on Data Mining, Maebashi City, Japan, 9–12 December 2002; pp. 306–313. [Google Scholar] [CrossRef][Green Version]
- Chinchor, N. MUC-4 Evaluation Metrics. In Proceedings of the 4th Conference on Message Understanding; Association for Computational Linguistics: Stroudsburg, PA, USA, 1992; pp. 22–29. [Google Scholar] [CrossRef][Green Version]
- Forman, G.; Scholz, M. Apples-to-Apples in Cross-Validation Studies: Pitfalls in Classifier Performance Measurement. SIGKDD Explor. Newsl.
**2010**, 12, 49–57. [Google Scholar] [CrossRef] - Platt, J.C. Sequential Minimal Optimization: A Fast Algorithm for Training Support Vector Machines; MIT Press: Cambridge, MA, USA, 1998. [Google Scholar]
- Robbins, H.; Monro, S. A Stochastic Approximation Method. Ann. Math. Statist.
**1951**, 22, 400–407. [Google Scholar] [CrossRef] - Chollet, F. Xception: Deep Learning With Depthwise Separable Convolutions. In Proceedings of the 2009 IEEE Conference on Computer Vision and Pattern Recognition, Miami, FL, USA, 20–25 June 2009; pp. 248–255. [Google Scholar]
- Melekhov, I.; Kannala, J.; Rahtu, E. Siamese network features for image matching. In Proceedings of the 2016 23rd International Conference on Pattern Recognition (ICPR), Cancun, Mexico, 4–8 December 2016; IEEE: Piscataway, NJ, USA, 2016; pp. 378–383. [Google Scholar]
- Szegedy, C.; Vanhoucke, V.; Ioffe, S.; Shlens, J.; Wojna, Z. Rethinking the Inception Architecture for Computer Vision. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, NV, USA, 26 June–1 July 2016; pp. 2818–2826. [Google Scholar]

**Figure 1.**Random Forest example where each tree classifies the new instance and the resulting class is decided by majority voting.

**Figure 6.**Histogram of all the used descriptors applied to a sample image. The vertical axis represents the number of occurrences of each texture unit normalized and the horizontal axis represents each of the texture units of the histograms. The descriptors are the ones that are part of D described at the beginning of Section 4.

**Figure 7.**Comparative of highest F1 of each iteration of the Subset Selection techniques. FSS stands for Forward Subset Selection and BSS stands for Backward Subset Selection. The numbers after each selection technique stand for the number of windows the descriptor has been applied to.

**Figure 8.**Times to classify an image with the different Subset selection methods on each level. The times on this Figure correspond to the average time that the classifier needs to classify an image using each descriptor on that level. The times of the BSS$8\times 8$ are not shown since its values are around 200 ms and distorts the plot.

**Table 1.**Forward Subset Selection of descriptors applied to the whole image (also known as FSS$4\times 4$). Level 1 uses only one descriptor. The following levels concatenate the best descriptor from the previous level to the rest of the descriptors. The algorithm stops on level 4 because the evaluation measure is not improved from level 3 to 4.

Descriptor | Level 1 | Level 2 | Level 3 | Level 4 |
---|---|---|---|---|

WLD + | WLD + BGC2 + | WLD + BGC2 + MTS + | ||

BGC1 | 0.66 (RF) | 0.931 (RF) | 0.937 (SVM) | 0.934 (SVM/RF) |

BGC2 | 0.489 (SVM) | 0.969 (SVM) | — | — |

BGC3 | 0.611 (SVM) | 0.937 (RC) | 0.929 (RF) | 0.929 (RF) |

GaborLBP | 0.671 (RF) | 0.931 (RF) | 0.903 (RF) | 0.906 (RF) |

GLCM | 0.811 (RF) | 0.951 (RF) | 0.903 (RF) | 0.963 (SVM) |

HOG | 0.84 (SVM) | 0.923 (SVM) | 0.923 (SVM) | 0.923 (SVM) |

LBP | 0.611 (RF) | 0.946 (RF) | 0.917 (SVM) | 0.923 (RF) |

LQP | 0.697 (RF) | 0.949 (SVM) | 0.954 (SVM) | 0.949 (SVM) |

LTP | 0.563 (RF) | 0.966 (SVM) | 0.969 (SVM) | 0.966 (SVM) |

MTS | 0.666 (KNN) | 0.966 (SVM) | 0.971 (SVM) | — |

STU1 | 0.746 (SVM) | 0.951 (SVM) | 0.951 (SVM) | 0.948 (SVM) |

STU2 | 0.74 (SVM) | 0.951 (SVM) | 0.957 (SVM) | 0.957 (SVM) |

WLD | 0.94 (RF) | — | — | — |

**Table 2.**Forward Subset Selection of descriptors applied to $4\times 4$ gridded image (also known as $FSS4\times 4$).

Descriptor | Level 1 | Level 2 | Level 3 |
---|---|---|---|

STU1 + | STU1 + WLD + | ||

BGC1 | 0.877 (SVM) | 0.908 (SVM-SGD) | 0.934 (SVM) |

BGC2 | 0.903 (SVM) | 0.931 (SVM) | 0.969 (SVM) |

BGC3 | 0.857 (SVM) | 0.906 (SVM) | 0.931 (SVM) |

GaborLBP | 0.834 (SVM) | 0.883 (SVM) | 0.903 (SVM) |

GLCM | 0.923 (RF) | 0.957 (SVM) | 0.966 (SVM) |

HOG | 0.846 (KNN) | 0.917 (SVM) | 0.94 (SVM) |

LBP | 0.874 (SVM) | 0.906 (SVM) | 0.929 (SVM) |

LQP | 0.911 (SVM) | 0.94 (SVM) | 0.969 (SVM) |

LTP | 0.889 (SVM) | 0.931 (SVM) | 0.96 (SVM) |

MTS | 0.909 (SVM) | 0.94 (SVM) | 0.96 (SVM) |

STU1 | 0.934 (SVM) | — | — |

STU2 | 0.914 (SVM) | 0.931 (SVM) | 0.96 (SVM) |

WLD | 0.931 (SVM) | 0.969 (SVM) | — |

**Table 3.**Forward Subset Selection of descriptors applied to $8\times 8$ gridded image (also known as $FSS8\times 8$).

Descriptor | Level 1 | Level 2 | Level 3 |
---|---|---|---|

WLD + | WLD + MTS + | ||

BGC1 | 0.845 (SVM) | 0.877 (SVM) | 0.897 (SVM) |

BGC2 | 0.911 (SVM) | 0.954 (SVM) | 0.957 (SVM) |

BGC3 | 0.843 (SVM) | 0.863 (RC) | 0.877 (SVM) |

GaborLBP | 0.783 (SVM) | 0.849 (Bagging) | 0.869 (Bagging) |

GLCM | 0.909 (SVM) | 0.92 (SVM) | 0.923 (SVM) |

HOG | 0.846 (KNN) | 0.923 (SVM) | 0.909 (SVM) |

LBP | 0.831 (SVM) | 0.886 (Bagging) | 0.889 (Bagging) |

LQP | 0.897 (SVM) | 0.929 (SVM) | 0.931 (SVM) |

LTP | 0.9 (SVM) | 0.951 (SVM) | 0.954 (SVM) |

MTS | 0.903 (SVM) | 0.96 (SVM) | — |

STU1 | 0.921 (SVM) | 0.94 (SVM) | 0.949 (SVM) |

STU2 | 0.917 (SVM) | 0.929 (SVM) | 0.937 (SVM) |

WLD | 0.94 (RF) | — | — |

**Table 4.**Backward Subset Selection of descriptors applied to $1\times 1$ gridded image (also known as BSS$1\times 1$). Level 1 uses all the descriptors in set D concatenated. Level 2 uses the concatenation of the descriptors in D without each of the descriptors. The following levels use the concatenation of the descriptors in D without the descriptor that makes score higher of the previous level.

Descriptor | Level 1 | Level 2 | Level 3 | Level 4 | Level 5 | Level 6 |
---|---|---|---|---|---|---|

D | D\ | D\ GaborLBP + | D\ GaborLBP + HOG + | D\ GaborLBP + HOG + BGC1 + | D\ GaborLBP + HOG + BGC1 + LBP + | |

BGC1 | 0.917 (SVM) | 0.92 (SVM) | 0.923 (SVM) | 0.929 (SVM) | — | — |

BGC2 | 0.914 (SVM) | 0.92 (SVM) | 0.92 (RF) | 0.937 (SVM) | 0.931 (SVM) | |

BGC3 | 0.917 (SVM) | 0.92 (SVM) | 0.929 (SVM) | 0.934 (SVM) | 0.937 (SVM) | |

LBP | 0.914 (SVM) | 0.92 (SVM) | 0.926 (RF) | 0.937 (SVM) | — | |

GaborLBP | 0.92 (SVM) | — | — | — | — | |

GLCM | 0.914 (SVM) | 0.914 (SVM) | 0.92 (SVM) | 0.929 (SVM) | 0.923 (SVM) | |

HOG | 0.903 (SVM) | 0.923 (SVM) | — | — | — | |

LQP | 0.917 (RF) | 0.92 (SVM) | 0.923 (SVM) | 0.923 (SVM) | 0.934 (SVM) | |

LTP | 0.909 (SVM) | 0.92 (SVM) | 0.92 (RF) | 0.926 (SVM) | 0.929 (SVM) | |

MTS | 0.914 (SVM) | 0.92 (SVM) | 0.923 (SVM) | 0.929 (SVM) | 0.931 (SVM) | |

STU1 | 0.917 (SVM) | 0.92 (SVM) | 0.923 (RF) | 0.923 (SVM) | 0.926 (SVM) | |

STU2 | 0.914 (SVM) | 0.92 (SVM) | 0.926 (RF) | 0.931 (SVM) | 0.926 (SVM) | |

WLD | 0.914 (SVM) | 0.891 (SVM) | 0.82 (RF) | 0.834 (RF) | 0.934 (RF) |

**Table 5.**Backward Subset Selection of descriptors applied to $4\times 4$ gridded image (also known as BSS$4\times 4$).

Descriptor | Level 1 | Level 2 | Level 3 | Level 4 |
---|---|---|---|---|

D | D \ | D \ LBP + | D \ LBP + STU1 + | |

BGC1 | 0.943 (SVM) | 0.95 (SVM) | 0.951 (SVM) | 0.949 (SVM) |

BGC2 | 0.946 (SVM) | 0.951 (SVM) | 0.954 (SVM) | |

BGC3 | 0.949 (SVM) | 0.95 (SVM) | 0.945 (SVM) | |

LBP | 0.951 (SVM) | — | — | |

GaborLBP | 0.946 (SVM) | 0.951 (RF) | 0.949 (RF) | |

GLCM | 0.937 (SVM) | 0.937 (SVM) | 0.94 (SVM) | |

HOG | 0.94 (SVM) | 0.943 (SVM) | 0.946 (SVM) | |

LQP | 0.946 (RF) | 0.946 (SVM) | 0.946 (SVM) | |

LTP | 0.946 (SVM) | 0.951 (SVM) | 0.949 (SVM) | |

MTS | 0.946 (SVM) | 0.951 (SVM) | 0.954 (SVM) | |

STU1 | 0.946 (SVM) | 0.954 (SVM) | — | |

STU2 | 0.949 (SVM) | 0.95 (SVM) | 0.949 (SVM) | |

WLD | 0.946 (SVM) | 0.946 (SVM) | 0.946 (SVM) |

**Table 6.**Backward Subset Selection of descriptors applied to $8\times 8$ gridded image (also known as BSS$8\times 8$).

Descriptor | Level 1 | Level 2 | Level 3 | Level 4 | Level 5 |
---|---|---|---|---|---|

D | D\ | D\ LBP + | D\ LBP + GaborLBP + | D\ LBP + GaborLBP + BGC3 + | |

BGC1 | 0.92 (SVM) | 0.914 (SVM) | 0.923 (SVM) | 0.943 (SVM) | 0.931 (SVM) |

BGC2 | 0.909 (SVM) | 0.929 (SVM) | 0.934 (SVM) | 0.943 (SVM) | |

BGC3 | 0.917 (SVM) | 0.926 (SVM) | 0.946 (SVM) | — | |

LBP | 0.926 (SVM) | — | — | — | |

GaborLBP | 0.92 (SVM) | 0.934 (RF) | — | — | |

GLCM | 0.894 (SVM) | 0.9 (SVM) | 0.917 (SVM) | 0.929 (SVM) | |

HOG | 0.903 (SVM) | 0.914 (SVM) | 0.934 (SVM) | 0.946 (SVM) | |

LQP | 0.906 (RF) | 0.917 (SVM) | 0.931 (SVM) | 0.934 (SVM) | |

LTP | 0.909 (SVM) | 0.914 (SVM) | 0.929 (SVM) | 0.934 (SVM) | |

MTS | 0.909 (SVM) | 0.937 (SVM) | 0.929 (SVM) | 0.946 (SVM) | |

STU1 | 0.906 (SVM) | 0.929 (SVM) | 0.934 (SVM) | 0.937 (SVM) | |

STU2 | 0.914 (SVM) | 0.929 (SVM) | 0.937 (SVM) | 0.943 (SVM) | |

WLD | 0.906 (SVM) | 0.923 (SVM) | 0.931 (SVM) | 0.943 (SVM) |

Method | F1 |
---|---|

Xception | 0.35 |

Siamese | 0.89 |

Our proposal (FSS$1\times 1$) | 0.97 |

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## Share and Cite

**MDPI and ACS Style**

Merino, I.; Azpiazu, J.; Remazeilles, A.; Sierra, B.
Histogram-Based Descriptor Subset Selection for Visual Recognition of Industrial Parts. *Appl. Sci.* **2020**, *10*, 3701.
https://doi.org/10.3390/app10113701

**AMA Style**

Merino I, Azpiazu J, Remazeilles A, Sierra B.
Histogram-Based Descriptor Subset Selection for Visual Recognition of Industrial Parts. *Applied Sciences*. 2020; 10(11):3701.
https://doi.org/10.3390/app10113701

**Chicago/Turabian Style**

Merino, Ibon, Jon Azpiazu, Anthony Remazeilles, and Basilio Sierra.
2020. "Histogram-Based Descriptor Subset Selection for Visual Recognition of Industrial Parts" *Applied Sciences* 10, no. 11: 3701.
https://doi.org/10.3390/app10113701