(35.175.191.168) 您好!臺灣時間:2019/10/15 01:41
字體大小: 字級放大   字級縮小   預設字形  
回查詢結果

詳目顯示:::

我願授權國圖
本論文永久網址: 
line
研究生:蔡緯豐
研究生(外文):Wei-Feng Tsai
論文名稱:基於深度神經網路的手勢辨識研究
論文名稱(外文):Hand Gesture Recognition Based on Deep Neural Network
指導教授:鍾鴻源莊堯棠
指導教授(外文):Hung-Yuan ChungYau-Tarng Juang
學位類別:碩士
校院名稱:國立中央大學
系所名稱:電機工程學系
學門:工程學門
學類:電資工程學類
論文出版年:2018
畢業學年度:106
語文別:中文
論文頁數:84
中文關鍵詞:手部偵測KCF追蹤CNN手勢辨識
外文關鍵詞:hand detectionKCF trackingCNNgesture recognition
相關次數:
  • 被引用被引用:0
  • 點閱點閱:83
  • 評分評分:系統版面圖檔系統版面圖檔系統版面圖檔系統版面圖檔系統版面圖檔
  • 下載下載:0
  • 收藏至我的研究室書目清單書目收藏:0
本文的目標是要實現使用網路攝影機即時追蹤影像範圍內的手部區域並且辨識手勢,應用於家電控制與人機互動等領域。我們首先利用膚色檢測和形態學處理分離影像,去除不必要的訊息,再利用背景相減法抓取手部的位置的區域ROI(Region Of Interest)。接著,為了避免雜訊影響到手部區塊,我們使用KCF(Kernelized Correlation Filters)演算法追蹤偵測到的手部區域ROI。最後將ROI的大小調整到100 * 120的大小,再將圖像輸入CNN (Convolutional Neural Networks)網路中進行多種手勢的辨識。接著重複上述追蹤和辨識的步驟達到即時的效果。本研究使用參考Alexnet和VGGnet網路的兩種架構進行訓練和比較,最後在訓練數據集中達到99.9%的辨識率,測試數據集有95.61%的辨識率。
The purposes of this paper are to achieve hand gesture recognition and tracking hand position in real time via web camera. First, using skin-color detect and morphological operations to remove unnecessary noise. Then use the background subtraction method to determine the ROI(Region Of Intereest) region of hand. After obtaining the hand region, Kernel Correlation Filters (KCF) algorithm is used to track the hand. Finally, the hand area is scaled to the size of 100 * 120, then the fixed size of the image input to our CNN (Convolutional Neural Networks) network for identification, in order to achieve the effect of identifying a variety of gestures. And repeat tracking and identification to achieve the real time performance.This research used two frameworks which referenced Alexnet and VGGnet for training and comparison. Finlly, a 99.9% recognition rate is achieved in the training data. The test data set has a recognition rate of 95.61%.
中文摘要 iv
英文摘要 v
致謝 vi
目錄 vii
圖目錄 x
表目錄 xiii
第一章 緒論 - 1 -
1-1 簡介 - 1 -
1-2 文獻回顧 - 2 -
1-3 研究動機與方法 - 3 -
1-4 主要貢獻 - 5 -
1-5 論文架構 - 5 -
第二章 系統描述 - 6 -
2-1 硬體 - 6 -
2-2 使用軟體 - 7 -
2-3 系統架構 - 9 -
第三章 手部偵測與追蹤 - 10 -
3-1 各種色彩空間介紹 - 11 -
3-1-1 RGB色彩空間 - 11 -
3-1-2 HSV色彩空間 - 11 -
3-1-3 YCbCr色彩空間 - 13 -
3-2 膚色切割法 - 13 -
3-3 形態學處理與平滑 - 16 -
3-3-1 侵蝕(Erosion) - 17 -
3-3-2 膨脹(Dilation) - 18 -
3-3-3 斷開(Opening) - 19 -
3-3-4 閉合(Closing) - 20 -
3-3-5 平滑處理 - 21 -
3-3-6 手部雜訊處理 - 22 -
3-3-7 背景相減法 - 24 -
3-4 手部追蹤 - 26 -
3-4-1 Linear regression - 26 -
3-4-2 Cyclic shift - 27 -
3-4-3 Kernal - 30 -
3-4-4 Fast detection - 31 -
3-4-5 Kernel Correlation - 32 -
3-4-6 追蹤流程 - 33 -
第四章 手勢辨識 - 35 -
4-1 卷積神經網路 - 36 -
4-1-1 卷積層(Convolutional layer) - 37 -
4-1-2 線性整流層(Rectified Linear Units layer, ReLU layer) - 40 -
4-1-3 池化層( Pooling Layer ) - 41 -
4-1-4 全連接層( Full connected layer ) - 42 -
4-2辨識架構 - 43 -
4-3訓練方法 - 46 -
第五章 實驗結果與討論 - 50 -
5-1實驗流程 - 51 -
5-2 訓練最加化 - 52 -
5-3 網路架構與結果 - 58 -
5-4 討論 - 64 -
第六章 結論與建議 - 66 -
參考文獻 - 67 -
[1] S. Suzuki, “Topological structural analysis of digitized binary images
by border following. Computer vision, graphics ”, and image processing, 30(1): 32-46 , 1985.
[2] J. Sklansky, “Finding the convex hull of a simple polygon. Pattern
Recognition Letters ”, 1(2): 79-83 , 1982.
[3] Srinivas Ganapathyraju, “ Hand gesture recognition using convexity hull defects to control an industrial robot ”, Instrumentation Control and Automation (ICA), 28-30 Aug, 2013.
[4] Microsoft, “Kinect for Windows”, http://www.kinectforwindows.org/,
2012.
[5] A. Memo, L. Minto, P. Zanuttigh, “Exploiting Silhouette Descriptors and Synthetic Data for Hand Gesture Recognition ”, 2015.
[6] Cem Keskin, Furkan Kıraç, Yunus Emre Kara, Lale Akarun, “Real time hand pose estimation using depth sensors ”, IEEE International Conference on Computer Vision Workshops, pp. 6-13 Nov, 2011.
[7] Leo Breiman , “Random Forests ”, Journal Machine Learning pp. 5-32 Oct, 2001.
[8] Trong-Nguyen Nguyen, Duc-Hoang Vo, Huu-Hung Huynh, Jean Meunier, “Geometry-based static hand gesture recognition using support vector machine”, Control Automation Robotics & Vision (ICARCV), 10-12 Dec, 2014.
[9] T. Starner, A. Pentland, “ Real-time american sign language recognition from video using hidden markov models ”, Motion-Based
Recognition. Springer Netherlands, 227-243, 1997.
[10] A. Krizhevsky, I. Sutskever , G. E. Hinton, “ Imagenet classification
with deep convolutional neural networks ”, Advances in neural information processing systems, 1097-1105 , 2012.
[11] ImageNet Large Scale Visual Recognition Challenge (ILSVRC) , “ http://www.image-net.org/challenges/LSVRC/ ”, 2010 – 2017.
[12] Mengmeng Han, Jiajun Chen, Ling Li, Yuchun Chang,“Visual Hand Gesture Recognition with Convolution Neural Network”, 21 July , 2016.
[13] P. KadewTraKuPong, R. Bowden,“An improved adaptive background mixture model for real-time tracking with shadow detection”, 2001.
[14] Z. Zivkovic, F. van der Heijden ,“Efficient adaptive density estimation per image pixel for the task of background subtraction”, 2006.
[15] João F. Henriques, Rui Caseiro, Pedro Martins, and Jorge Batista, “High-speed tracking with kernelized correlation filters”, Pattern Analysis and Machine Intelligence, IEEE Transactions on, vol. 37, no. 3, pp. 583–596, 2015.
[16] K. Simonyan and A. Zisserman,“Very deep convolutional networks for large-scale image recognition ” , arXiv preprint arXiv:1409.1556, 2014.
[17] OpenCV,“https://opencv.org/ ”, 2000
[18] Tensorflow,“https://www.tensorflow.org/ ”, 2015
[19] Youlian Zhu, Cheng Huang and Jiajun Chen, “Face Detection Method Based on Multi-feature Fusion in YCbCr Color Space ”, International Congress on Image and Signal Processing, pp. 1249-1252, Oct. 2012.
[20] C. Tomasi ; R. Manduchi,“Bilateral filtering for gray and color images”, Computer Vision, Sixth International Conference on , pp. 7-7 Jan , 1998.
[21] Maycas Nadal, Catlos,“Input-Output Kernel Regression applied to proteinprotein interaction network inferenc”, PhD Thesis , 2015.
[22] Yan LeCun“LeNet-5, convolutional neural networks ”, http: //yann.lecun.com/exdb/lenet/, 1998.
[23] Sebastian Ruder, “ An overview of gradient descent optimization
Algorithms ” , 2016.
[24] Ning Qian, “On the momentum term in gradient descent learning algorithms. ”, Neural networks :the official journal of the International Neural Network Society, 12(1):145–151 , 1999.
[25] Yurii Nesterov, “A method for unconstrained convex minimization problem with the rate of convergence o(1/k2) ”, Doklady ANSSSR (translated as Soviet.Math.Docl.), 269:543–547 , 1993.
[26] John Duchi, Elad Hazan, Yoram Singer ,“ Adaptive Subgradient Methods for Online Learning and Stochastic Optimization ” , Journal of Machine Learning Research, 2121–2159, 2011.
[27] Matthew D. Zeiler ,“ ADADELTA: An Adaptive Learning Rate Method ” , 2012.
[28] Diederik P. Kingma , Jimmy Lei Ba , “ Adam: a Method for Stochastic Optimization ”, InternationalConference on Learning Representations, pp. 1–13 , 2015.
[29] Srivastava, N., Hinton, G., Krizhevsky, A., Sutskever, I., & Salakhutdinov, R , “ Dropout: A simple way to prevent neural networks from overfitting. ”, The Journal of Machine Learning Research, 15(1), 1929-1958 , 2014
[30] Geoffrey E. Hinton, Nitish Srivastava, Alex Krizhevsky, Ilya Sutskever, Ruslan R. Salakhutdinov , “ Improving neural networks by preventing co-adaptation of feature detectors ” , 2012.
連結至畢業學校之論文網頁點我開啟連結
註: 此連結為研究生畢業學校所提供,不一定有電子全文可供下載,若連結有誤,請點選上方之〝勘誤回報〞功能,我們會盡快修正,謝謝!
QRCODE
 
 
 
 
 
                                                                                                                                                                                                                                                                                                                                                                                                               
第一頁 上一頁 下一頁 最後一頁 top
系統版面圖檔 系統版面圖檔