現代的人，使用電腦或行動裝置上網已經是每天習慣要做的事，瀏覽的網站從一般的內容型網站、社群網站、影音媒體網站到電子商務型網站都有，應該有人碰過，進到某個網站就發現頁面錯誤或是呈現的網頁內容是壞掉的，可能少了一張圖片、圖片與文字對不起來或是某個區塊跑到了不該出現的地方，這樣的狀況在業界稱之為跑版；相信建置網站的開發團隊都極不願意把這些跑版的資訊呈現在使用者的眼前，這樣不僅可能會流失網站的流量，最重要的是讓自己網站的品質受到了質疑與傷害；本研究主要在探討使用圖片/影像辯識的方法，對網站頁面轉成的圖片進行辨識是否有跑版的問題發生；實驗中使用深度學習在影像辨識領域表現得最好的卷積神經網路演算法，搭配圖片數量、圖片尺寸、訓練回合數、卷積層數等變因進行訓練，根據本研究實驗得到的結果顯示，若各變因有適當的
調整，則所獲得的準確率及混淆矩陣分類正確性都會獲得良好的改善。

Modern people, using computers or mobile devices to surf the Internet is a habit
to do every day. The websites browsed are from general content sites, community sites,
video sites to e-commerce sites. In some cases, when we visit a website, some webpage
layout is incorrect or the content of the presented webpage is broken. There may be a
missing picture, image and non-matched text, or a content block has moved to a place
where it should not appear. This situation is called "broken layout" within the industry.
It is true that the development team that built the website is very reluctant to show
the "broken layout" to end users, so that not only may the website traffic be lost, but
also the degradation for the quality of the website. Users would have doubt about the
quality and hurt brand.
This research is mainly to explore the method of using image identification to
identify whether the image converted from the website page has a "broken layout"
problem; deep learning is used in the experiment since it performs well in the field of
image identification. The neural network algorithm is trained with different factors such
as the number of training pictures, the size of the pictures, the number of training
iterations, and the number of convolutional layers. According to the results of this
research, if the various factors are adjusted appropriately, the accuracy rate obtained
and the confusion matrix classification accuracy will be improved.

【英文文獻】
Behnke, S. (2003) . Hierarchical neural networks for image interpretation. Springer. 2766
Cireşan, D. C., Meier, U., Gambardella, L. M., & Schmidhuber, J. (2010) . Deep, big, simple neural nets for handwritten digit recognition. Neural Computation, 22 (12) , 3207-3220.
Fukushima, K. (1980) . Neocognitron: A self-organizing neural network model for a mechanism of pattern recognition unaffected by shift in position. Biological cybernetics, 36 (4) , 193-202.
Hinton, G. E., & Salakhutdinov, R. R. (2006) . Reducing the dimensionality of data with neural networks. science, 313 (5786) , 504-507.
Hubel, D. H. (1962) . Receptive fields, binocular interaction and functional architecture in the cat's visual cortex. The Journal of physiology, 160, 106-154.
Hubel, D. H., & Wiesel, T. N. (1968) . Receptive fields and functional architecture of monkey striate cortex. The Journal of physiology, 195 (1) , 215-243.
Le, Q. V. (2013) . Building high-level features using large scale unsupervised learning. Paper presented at the 2013 IEEE international conference on acoustics, speech and signal processing.
LeCun, Y., Bengio, Y., & Hinton, G. (2015) . Deep learning. nature, 521 (7553) , 436-444.
LeCun, Y., Boser, B., Denker, J. S., Henderson, D., Howard, R. E., Hubbard, W., & Jackel, L. D. (1989) . Backpropagation applied to handwritten zip code recognition. Neural Computation, 1 (4) , 541-551.
LeCun, Y., Bottou, L., Bengio, Y., & Haffner, P. (1998) . Gradient-based learning applied to document recognition. Proceedings of the IEEE, 86 (11) , 2278-2324.
Markoff, J. (2012) . How many computers to identify a cat? 16,000. New York Times, 25.
Schmidhuber, J. (1992) . Learning complex, extended sequences using the principle of history compression. Neural Computation, 4 (2) , 234-242.
Schmidhuber, J. (2015) . Deep learning in neural networks: An overview. Neural networks, 61, 85-117.
Taigman, Y., Yang, M., Ranzato, M. A., & Wolf, L. (2014) . Deepface: Closing the gap to human-level performance in face verification. Paper presented at the Proceedings of the IEEE conference on computer vision and pattern recognition.
Werbos, P. (1974) . Beyond regression:" new tools for prediction and analysis in the behavioral sciences. Ph. D. dissertation, Harvard University.

【中文文獻】
郭柏宏 (2015) 。基於深度學習之靜態影像超解析度技術。國立成功大學電機工程研究所博碩士論文。

【書籍與網站】
Brandon (2016) . End to End Machine Learning - How do Convolutional Neural Networks work? Retrieved June, 2020 from https://e2eml.school/how_convolutional_neural_networks_work.html
CH.Tseng (2017, September 12) 。初探卷積神經網路。Retrieved June, 2020 from: https://chtseng.wordpress.com/2017/09/12/初探卷積神經網路/
Cinnamon AI Taiwan (2019, June 5) 。深度學習：CNN原理。Retrieved June, 2020 from: https://medium.com/@CinnamonAITaiwan/深度學習-cnn原理-keras實現-432fd9ea4935
Machine Learning Notebook (2017, April 07) . Convolutional Neural Networks - Basics. Retrieved June, 2020 from: https://mlnotebook.github.io/post/CNN1/
Steven Shen (2018, January 2) 。入門深度學習—2。Retrieved June, 2020 from: https://medium.com/@syshen/入門深度學習-2-d694cad7d1e5
三津村直貴 (2018) 。圖解AI人工智慧大未來：關於人工智慧一定要懂的96件事。臺灣：旗標出版社。
全球資訊網協會 (World Wide Web Consortium，W3C) 簡介 (2020.06) 取自：https://www.w3.org/
林大貴 (2017) 。TensorFlow+Keras 深度學習人工智慧實務應用。臺灣：博碩出版社。