本論文藉著改良三元損失（Triplet loss）並應用到深度學習神經網路中藉以提取皮革影像紋理特徵來辨識皮革紋路。在這項研究中，我們提出一種新的三元損失，改變原本三元損失以樣本在歐式集合空間中以距離大小引導卷積神經網路的特徵學習，而改以增強樣本與正樣本間的相似性並降低樣本與負樣本之間的相似性為訴求來設計三元損失函數，進而引導卷積神經網路的學習，在特徵學習成功後，我們利用SVM進行皮革紋路的辨識，而非直接以神經網路進行辨識，如此做法可以避免新皮革加入資料庫時必須重新訓練整個包含特徵抽取以及皮革辨識的神經網路的問題，我們只需重新訓練SVM即可，無需重新訓練特徵抽取的網路，透過實驗結果與比較，我們所提出的方法對紋理識別是有不錯的改善效果。

In this thesis, the triplet Loss is refined and applied to the deep learning neural network to extract texture features of leather images to identify leather texture patterns. In this study, we put forward a new kind of triplet loss that changes the way of the original triplet loss which guide the feature learning of a convolutional neural network (CNN) with the sample distances in Euclidean space to another way which enhances the similarity between intra-class samples and reduces the similarity between inter-class samples. After successfully learning the effective features with the CNN, we use a support vector machine (SVM), instead of using the CNN directly, to recognize the leather texture patterns. This way can avoid the retraining of the CNN for both the feature learning and the leather texture recognition whenever any new leather is added into the database. Instead, our method only needs to retrain the SVM for recognizing the newly added leather. Through a number of experiments and comparisons we have made, the results show that our proposed method achieves satisfactory recognition performance.

誌謝 I
摘要 II
Abstract III
圖目錄 VI
表目錄 VII
第 1 章 緒論 1
1.1 研究動機 1
1.2 系統流程 3
第 2 章 文獻探討 5
2.1 卷積神經網路(Convolutional Neural Network) 5
2.2 三元損失（Triplet Loss） 6
2.3 GoogLeNet 7
第 3 章 三元卷積神經網路特徵抽取與SVM辨識 8
3.1 填充方式（Padding） 9
3.2 卷積神經網路架構 10
3.3 改良型三元度量損失（Improved Triplet Metric Loss） 11
第 4 章 實驗結果與討論 16
4.1 資料處理 16
4.2 樣本訓練與測試方式及辨識率計算 17
4.3 效能評估與比較 18
4.3.1 較小資料量之全局最大∕平均池化結果比較 19
4.3.2 較大資料量不同分類器結果比較 20
4.3.3 較大量外部資料辨識實驗 22
4.3.4 損失函數α值測試 23
4.3.5 濾波器層數測試 24
4.3.6 皮革切割大小測試 25
4.3.7 較大量內部與外部資料辨識比較與分析 25
第 5 章 結論與未來研究方向 29
參考文獻 30

[1]C. Szegedy et al., "Going deeper with convolutions," 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Boston, MA, 2015, pp. 1-9.
[2]L. K. Seng and T. Guniawan, "An Experimental Study on the Use of Visual Texture for Wood Identification Using a Novel Convolutional Neural Network Layer," 2018 8th IEEE International Conference on Control System, Computing and Engineering (ICCSCE), Penang, Malaysia, 2018, pp. 156-159.
[3]T. Khurana, K. Mahajan, C. Arora and A. Rai, "Exploiting Texture Cues for Clothing Parsing in Fashion Images," 2018 25th IEEE International Conference on Image Processing (ICIP), Athens, 2018, pp. 2102-2106.
[4]G. Zhu, B. Li, S. Hong and B. Mao, "Texture Recognition and Classification Based on Deep Learning," 2018 Sixth International Conference on Advanced Cloud and Big Data (CBD), Lanzhou, 2018, pp. 344-348.
[5]M. Cimpoi, S. Maji and A. Vedaldi, "Deep filter banks for texture recognition and segmentation," 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Boston, MA, 2015, pp. 3828-3836.
[6]M. Jawahar, N. K. C. Babu and K. Vani, "Leather texture classification using wavelet feature extraction technique," 2014 IEEE International Conference on Computational Intelligence and Computing Research, Coimbatore, 2014, pp. 1-4.
[7]Xin Wang and N. D. Georganas, "GLCM texture based fractal method for evaluating fabric surface roughness," 2009 Canadian Conference on Electrical and Computer Engineering, St. John's, NL, 2009, pp. 104-107.
[8]Q. Chen and E. Agu, "Exploring Statistical GLCM Texture Features for Classifying Food Images," 2015 International Conference on Healthcare Informatics, Dallas, TX, 2015, pp. 453-453.
[9]F. Schroff, D. Kalenichenko and J. Philbin, "FaceNet: A Unified Embedding for Face Recognition and Clustering," 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2015, pp. 815-823.
[10]H. Liu and W. Huang, "Body structure based triplet Convolutional Neural Network for person re-identification," 2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), New Orleans, LA, 2017, pp. 1772-1776.
[11]W. Wan and H. J. Lee, "FaceNet Based Face Sketch Recognition," 2017 International Conference on Computational Science and Computational Intelligence (CSCI), Las Vegas, NV, 2017, pp. 432-436.
[12]H. Zhan, B. Shi and A. C. Kot, "Street-to-shop shoe retrieval with multi-scale viewpoint invariant triplet network," 2017 IEEE International Conference on Image Processing (ICIP), Beijing, 2017, pp. 1102-1106.
[13]M. Lin, Q. Chen, S. Yan, "Network in network", CoRR, 2013.
[14]C. Szegedy et al., "Rethinking the Inception Architecture for Computer Vision" 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016, pp. 2818-2826.
[15]F. Mirzapour and H. Ghassemian, "Using GLCM and Gabor filters for classification of PAN images," 2013 21st Iranian Conference on Electrical Engineering (ICEE), Mashhad, 2013, pp. 1-6.
[16]S. Dongcheng, C. Fang and D. Guangyi, "Facial Expression Recognition Based on Gabor Wavelet Phase Features," 2013 Seventh International Conference on Image and Graphics, Qingdao, 2013, pp. 520-523.
[17]Z. Dai-xian, S. Zhe and W. Jing, "Face recognition method combined with gamma transform and Gabor transform," 2015 IEEE International Conference on Signal Processing, Communications and Computing (ICSPCC), Ningbo, 2015, pp. 1-4.
[18]Y. Lecun, L. Bottou, Y. Bengio and P. Haffner, "Gradient-based learning applied to document recognition," in Proceedings of the IEEE, vol. 86, no. 11, pp. 2278-2324, Nov. 1998.
[19]A. Krizhevsky, I. Sutskever and G. E. Hinton, "ImageNet Classification with Deep Convolutional Neural Networks." 2012 Proceedings of the 25th International Conference on Neural Information Processing Systems, p.1097-1105, December 03-06, 2012, Lake Tahoe, Nevada.
[20]K. Simonyan and A. Zisserman, "Very Deep Convolutional Networks for Large-Scale Image Recognition" In ICLR 2015.
[21]K. He, X. Zhang , S. Ren and J. Sun, "Deep Residual Learning for Image Recognition" 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016, pp. 770-778.
[22]C. Cortes and V. Vapnik, “Support-Vector Networks” Machine Learning, 20:273–297, November 1995.