本論文提出了基於深度主動學習之表面瑕疵切割方法，是一種藉由將主動學習機制添加至深度學習網路，使其可以透過人工干預進行網路優化的瑕疵檢測技術。
主動學習方法旨在設計一套演算流程，透過分析尚未進行標註的樣本經模型計算出的各項參數，並對其進行各項計算後，決定欲從未標註資料集中挑選哪些候選樣本進行人工標註。多數的主動學習算法會根據候選影像的不確定性及多樣性進而決定優先序，而其最終目標主要為兩點，分別為1)僅對被挑選出的資料而非整個未標注資料集進行人工標註，以降低人工標註成本2)使機制針對利於優化網路的對象進行挑選，因此網路得以使用較低的資料量獲得不錯的成果。
而本論文主要針對物件表面瑕疵進行實驗，利用影像中元素皆為同材質的特性，對未標註影像的深度特徵進行比較，比較的對象分為：同張候選影像，不同位置的深度特徵、不同張候選影像之間的相似度比較。
在實驗結果方面，本論文使用Crack-Forest Dataset進行測試，由於模擬實際應用時容易存在正負樣本不平衡的情況，本論文將該資料集之影像進行切割並重整後進行實驗分析。實驗成果證明，添加主動學習機制可有效的於前幾輪挑選出較多具瑕疵的樣本，使精準度可以較快提升。雖然由於資料不平衡及瑕疵總量不足的問題使切割精準度不高，但對於資料初步篩選並預測資料瑕疵位置仍具一定能力。

This study proposed a surface defect segmentation method based on deep active learning. By adding active learning mechanism to deep learning architecture, the model can be optimized through manual intervention.
Active learning is a mechanism to design a set of computing processes to determine which unlabeled image is selected by analyzing the data computed by the model. Most active learning algorithms determine the priority based on the uncertainty and diversity of the candidate images. Their ultimate goals are mainly twofold. One is to annotate the selected data instead of the entire unlabeled dataset to reduce the labor cost. The other is to make the model select objects that are conducive to the optimization of the model so that the model can obtain good results with a small amount of data.
This study makes experiments on the surface defects of the object. Take advantage of the characteristics of all elements in images is in the same material, the objects of comparison are divided into two parts. First, compare the deep features of different positions of the same candidate. Second, compare the similarity between different candidates.
This study conducts experiments on the Crack-Forest Dataset to simulate the imbalance of positive and negative samples in practical applications, we cropped and reformates the images of the dataset. Experimental results show that adding active learning mechanism can effectively select more defective samples in the first few rounds so that the accuracy can be improved quickly. Although the accuracy is not high due to the imbalance of data and insufficient defects, it still has a certain ability for preliminary filtrating the data and predicting the location of defects

[1] T. B. Sheridan and W. L. Verplank, "Human and computer control of undersea teleoperators," Massachusetts Inst of Tech Cambridge Man-Machine Systems Lab, 1978.
[2] Y. LeCun, L. Bottou, Y. Bengio, and P. Haffner, "Gradient-based learning applied to document recognition," Proceedings of the IEEE, vol. 86, no. 11, pp. 2278-2324, 1998.
[3] M.-C. Popescu, V. E. Balas, L. Perescu-Popescu, and N. Mastorakis, "Multilayer perceptron and neural networks," WSEAS Transactions on Circuits and Systems, vol. 8, no. 7, pp. 579-588, 2009.
[4] J. Deng, W. Dong, R. Socher, L.-J. Li, K. Li, and L. Fei-Fei, "Imagenet: A large-scale hierarchical image database," in 2009 IEEE conference on computer vision and pattern recognition, 2009: Ieee, pp. 248-255.
[5] A. Krizhevsky, I. Sutskever, and G. E. Hinton, "Imagenet classification with deep convolutional neural networks," Advances in neural information processing systems, vol. 25, pp. 1097-1105, 2012.
[6] K. Simonyan and A. Zisserman, "Very deep convolutional networks for large-scale image recognition," arXiv preprint arXiv:1409.1556, 2014.
[7] C. Szegedy et al., "Going deeper with convolutions," in Proceedings of the IEEE conference on computer vision and pattern recognition, 2015, pp. 1-9.
[8] K. He, X. Zhang, S. Ren, and J. Sun, "Deep residual learning for image recognition," in Proceedings of the IEEE conference on computer vision and pattern recognition, 2016, pp. 770-778.
[9] J. Long, E. Shelhamer, and T. Darrell, "Fully convolutional networks for semantic segmentation," in Proceedings of the IEEE conference on computer vision and pattern recognition, 2015, pp. 3431-3440.
[10] O. Ronneberger, P. Fischer, and T. Brox, "U-net: Convolutional networks for biomedical image segmentation," in International Conference on Medical image computing and computer-assisted intervention, 2015: Springer, pp. 234-241.
[11] V. Badrinarayanan, A. Kendall, and R. Cipolla, "Segnet: A deep convolutional encoder-decoder architecture for image segmentation," IEEE transactions on pattern analysis and machine intelligence, vol. 39, no. 12, pp. 2481-2495, 2017.
[12] L.-C. Chen, G. Papandreou, I. Kokkinos, K. Murphy, and A. L. Yuille, "Semantic image segmentation with deep convolutional nets and fully connected crfs," arXiv preprint arXiv:1412.7062, 2014.
[13] L.-C. Chen, G. Papandreou, I. Kokkinos, K. Murphy, and A. L. Yuille, "Deeplab: Semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected crfs," IEEE transactions on pattern analysis and machine intelligence, vol. 40, no. 4, pp. 834-848, 2017.
[14] K. He, X. Zhang, S. Ren, and J. Sun, "Spatial pyramid pooling in deep convolutional networks for visual recognition," IEEE transactions on pattern analysis and machine intelligence, vol. 37, no. 9, pp. 1904-1916, 2015.
[15] L.-C. Chen, G. Papandreou, F. Schroff, and H. Adam, "Rethinking atrous convolution for semantic image segmentation," arXiv preprint arXiv:1706.05587, 2017.
[16] L.-C. Chen, Y. Zhu, G. Papandreou, F. Schroff, and H. Adam, "Encoder-decoder with atrous separable convolution for semantic image segmentation," in Proceedings of the European conference on computer vision (ECCV), 2018, pp. 801-818.
[17] Z. Zhou, J. Shin, L. Zhang, S. Gurudu, M. Gotway, and J. Liang, "Fine-tuning convolutional neural networks for biomedical image analysis: actively and incrementally," in Proceedings of the IEEE conference on computer vision and pattern recognition, 2017, pp. 7340-7351.
[18] Y. Siddiqui, J. Valentin, and M. Nießner, "Viewal: Active learning with viewpoint entropy for semantic segmentation," in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2020, pp. 9433-9443.
[19] M. Van den Bergh, X. Boix, G. Roig, B. de Capitani, and L. Van Gool, "Seeds: Superpixels extracted via energy-driven sampling," in European conference on computer vision, 2012: Springer, pp. 13-26.
[20] F. Iandola, M. Moskewicz, S. Karayev, R. Girshick, T. Darrell, and K. Keutzer, "Densenet: Implementing efficient convnet descriptor pyramids," arXiv preprint arXiv:1404.1869, 2014.
[21] T. Chen, S. Kornblith, M. Norouzi, and G. Hinton, "A simple framework for contrastive learning of visual representations," in International conference on machine learning, 2020: PMLR, pp. 1597-1607.
[22] M. Arjovsky, S. Chintala, and L. Bottou, "Wasserstein generative adversarial networks," in International conference on machine learning, 2017: PMLR, pp. 214-223.
[23] M. Cuturi, "Sinkhorn distances: Lightspeed computation of optimal transport," Advances in neural information processing systems, vol. 26, pp. 2292-2300, 2013.
[24] J. Han, P. Luo, and X. Wang, "Deep self-learning from noisy labels," in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2019, pp. 5138-5147.
[25] Y. Shi, L. Cui, Z. Qi, F. Meng, and Z. Chen, "Automatic road crack detection using random structured forests," IEEE Transactions on Intelligent Transportation Systems, vol. 17, no. 12, pp. 3434-3445, 2016.