1. NCCU Academic Hub
  2. Publications
  3. College of Informatics
  4. Theses
  5. Item

Publications-Theses

Article View/Open

Publication Export

Google ScholarTM

NCCU Library

Citation Infomation

Related Publications in TAIR

題名 延展式區域三元化圖形特徵描述子之比例式降維法
Commensurate Dimensionality Reduction in Extended Local Ternary Patterns
作者 余浩瑋
貢獻者 廖文宏
余浩瑋
關鍵詞 特徵描述
區域二元化圖形
延展式區域三元化圖形
比例式降維法
模糊理論
紋理影像辨識
feature descriptor
local binary pattern
extend local ternary pattern
commensurate dimensionality reduction
fuzzy logic
texture classification
日期 2012
上傳時間 3-Mar-2014 15:38:56 (UTC+8)
摘要 區域二/三元化樣式與其各種變型被廣泛應用於物件辨識中的特徵描述，然而現有的區域特徵描述方式，普遍存在適用時機的問題，也就是針對不同類型的圖像資料庫，必須選用符合該圖片性質的描述法，方能達到較佳的辨識效果，舉例而言，處理材質影像時多使用uniform pattern，而進行人臉偵測或表情辨識時則多採用一般型的區域二/三元化樣式。

本論文的目標是建構一個通用型的區域三元化樣式，使其一體適用於各類圖型辨識的任務，我們以延展式區域三元化樣式(Extended local ternary patterns, ELTP)為基礎，探討各種降維演算法的結合機制，並提出可行的樣式定義方法，我們針對ETLP 中的uniform pattern定義重新思考，藉由大規模實驗與統計，探討各類uniform pattern 的從屬關係與出現比例，並依據比例原則，在降維階段分配適當之維度，稱之為比例式降維法。

本論文針對比例式降維後的ELTP之抗噪性、描述力與通用性進行深度的分析與廣泛的實驗，以驗證此類圖像描述方法之效能。此外，由於特徵樣式的編碼在某些情況下，容易受到雜訊影響而產生變化，因此我們亦提出了結合模糊理論的方式加以改善。以上改良方式之效果，都在各式實驗中獲得驗證。
Local binary/ternary pattern and its derivatives have been widely employed to represent low-level features in many pattern recognition tasks. However, existing local descriptors fail to achieve universal applicability in the sense that specific types of local binary patterns are better suited for certain collections of images. For example, uniform local binary patterns are preferred when dealing with textures, while regular local binary/ternary patterns are adopted for face detection and facial expression recognition.

This thesis proposes a universally applicable local descriptor based on the extended local ternary pattern (ELTP) to address the above concern. We exploit the feasibility of combining dimensionality reduction techniques to derive a novel local descriptor that is suitable for all kinds of object recognition applications. Specifically, we investigate all possible definitions of uniform patterns under ternary encoding scheme and study their properties. This enables us to devise a dimensionality assignment algorithm in which the allocated dimension is proportional to the appearance rate of the corresponding pattern group.

The newly defined extend local ternary pattern using commensurate dimensionality reduction (ELTP-CDR) technique has been extensively tested to analyze its universality, discriminability, and noise sensitivity. To further enhance the robustness of this class of local descriptor, we also incorporate fuzzy logic to derive fuzzy representations of the original ELTP. The efficacy of the proposed methods has been validated through several experiments targeted at texture recognition tasks.
參考文獻 [1] T. Ojala, M. Pietikainen, and T. Maenpaa, “Multiresolution Gray-Scale and Rotation Invariant Texture Classification with Local Binary Patterns,” IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 24, no. 7, pp. 971-987, July 2002.
[2] D. Lowe, “Distinctive Image Features from Scale-Invariant Keypoints,” Int`l J. Computer Vision, vol. 2, no. 60, pp. 91-110, 2004.
[3] H. Bay, A. Ess, T. Tuytelaars, L. V. Gool, “SURF: Speeded Up Robust Features”, Computer Vision and Image Understanding (CVIU), Vol. 110, No. 3, pp. 346--359, 2008.
[4] A. Alahi, R. Ortiz, P. Vandergheynst, “FREAK: Fast Retina Keypoint”, IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2012.
[5] S. Leutenegger, M. Chli, R. Y. Siegwart, “BRISK: Binary Robust Invariant Scalable Keypoints”, IEEE International Conference on Computer Vision (ICCV), 2011.
[6] W. H. Liao, “Region Description Using Extended Local Ternary Patterns”, Proceedings of the 20th International Conference on Pattern Recognition, pp. 1003-1006, 2010.
[7] X. Tan and B. Triggs. “Enhanced local texture feature sets for face recognition under difficult lighting conditions”. In Analysis and Modeling of Faces and Gestures, volume 4778 of LNCS, pages 168–182. Springer, 2007.
[8] A. Shobeirinejad and Y. S. Gao, “Gender Classification Using Interlaced Derivative Patterns“, Proceedings of the 20th International Conference on Pattern Recognition, pp. 1509-1512, 2010.
[9] T. Ojala, M. Pietikainen and T. Maenpaa, ”Multi-resolution Gray-scale and Rotation Invariant Texture Classification with Local Binary Patterns”, IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 24(7), pp.971-987. July 2002.
[10] M. Heikkilä, M. Pietikäinen and C. Schmid, “Description of Interest Regions with Center-Symmetric Local Binary Patterns”, Computer Vision, Graphics and Image Processing, Lecture Notes in Computer Science, 2006.
[11] T. Ahonen, M. Pietikainen, “Soft Histograms for Local Binary Patterns”, in Proc. Fin. Signal Process. Symp., Oulu, Finland, 2007.
[12] J. C. Dunn, “A Fuzzy Relative of the ISODATA Process and Its Use in Detecting Compact Well-Separated Clusters”, Journal of Cybernetics 3: 32-57, 1973
[13] R. Krishnapuram, A. Joshi, L. Yi, “A Fuzzy Relative of the k-Medoids Algorithm with Application to Web Document and Snippet Clustering”, Snippet Clustering, in Proc. IEEE Intl. Conf. Fuzzy Systems - FUZZIEEE99, Korea, 1999.
[14] T. Ahonen, A. Hadid, and M. Pietikainen, “Face Description with Local Binary Patterns: Application to Face Recognition”, IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 28, no. 12, pp.2037-2041, Dec. 2006.
[15] N. P. Doshi, G. Schaefer, “A Comprehensive Benchmark of Local Binary Pattern Algorithms for Texture Retrieval”, International Conference on Pattern Recognition (ICPR), 2012.
[16] Everingham, M., Van~Gool, L., Williams, C. K. I., Winn, J., Zisserman, A., “The PASCAL Visual Object Classes Challenge 2012 (VOC2012) Results”, http://www.pascal-network.org/challenges/VOC/voc2012/workshop/index.html.
描述 碩士
國立政治大學
資訊科學學系
98753030
101
資料來源 http://thesis.lib.nccu.edu.tw/record/#G0098753030
資料類型 thesis
dc.contributor.advisor 廖文宏zh_TW
dc.contributor.author (Authors) 余浩瑋zh_TW
dc.creator (作者) 余浩瑋zh_TW
dc.date (日期) 2012en_US
dc.date.accessioned 3-Mar-2014 15:38:56 (UTC+8)-
dc.date.available 3-Mar-2014 15:38:56 (UTC+8)-
dc.date.issued (上傳時間) 3-Mar-2014 15:38:56 (UTC+8)-
dc.identifier (Other Identifiers) G0098753030en_US
dc.identifier.uri (URI) http://nccur.lib.nccu.edu.tw/handle/140.119/64373-
dc.description (描述) 碩士zh_TW
dc.description (描述) 國立政治大學zh_TW
dc.description (描述) 資訊科學學系zh_TW
dc.description (描述) 98753030zh_TW
dc.description (描述) 101zh_TW
dc.description.abstract (摘要) 區域二/三元化樣式與其各種變型被廣泛應用於物件辨識中的特徵描述，然而現有的區域特徵描述方式，普遍存在適用時機的問題，也就是針對不同類型的圖像資料庫，必須選用符合該圖片性質的描述法，方能達到較佳的辨識效果，舉例而言，處理材質影像時多使用uniform pattern，而進行人臉偵測或表情辨識時則多採用一般型的區域二/三元化樣式。

本論文的目標是建構一個通用型的區域三元化樣式，使其一體適用於各類圖型辨識的任務，我們以延展式區域三元化樣式(Extended local ternary patterns, ELTP)為基礎，探討各種降維演算法的結合機制，並提出可行的樣式定義方法，我們針對ETLP 中的uniform pattern定義重新思考，藉由大規模實驗與統計，探討各類uniform pattern 的從屬關係與出現比例，並依據比例原則，在降維階段分配適當之維度，稱之為比例式降維法。

本論文針對比例式降維後的ELTP之抗噪性、描述力與通用性進行深度的分析與廣泛的實驗，以驗證此類圖像描述方法之效能。此外，由於特徵樣式的編碼在某些情況下，容易受到雜訊影響而產生變化，因此我們亦提出了結合模糊理論的方式加以改善。以上改良方式之效果，都在各式實驗中獲得驗證。		zh_TW
dc.description.abstract (摘要) Local binary/ternary pattern and its derivatives have been widely employed to represent low-level features in many pattern recognition tasks. However, existing local descriptors fail to achieve universal applicability in the sense that specific types of local binary patterns are better suited for certain collections of images. For example, uniform local binary patterns are preferred when dealing with textures, while regular local binary/ternary patterns are adopted for face detection and facial expression recognition.

This thesis proposes a universally applicable local descriptor based on the extended local ternary pattern (ELTP) to address the above concern. We exploit the feasibility of combining dimensionality reduction techniques to derive a novel local descriptor that is suitable for all kinds of object recognition applications. Specifically, we investigate all possible definitions of uniform patterns under ternary encoding scheme and study their properties. This enables us to devise a dimensionality assignment algorithm in which the allocated dimension is proportional to the appearance rate of the corresponding pattern group.

The newly defined extend local ternary pattern using commensurate dimensionality reduction (ELTP-CDR) technique has been extensively tested to analyze its universality, discriminability, and noise sensitivity. To further enhance the robustness of this class of local descriptor, we also incorporate fuzzy logic to derive fuzzy representations of the original ELTP. The efficacy of the proposed methods has been validated through several experiments targeted at texture recognition tasks.		en_US
dc.description.tableofcontents 第一章 緒論 1
1.1 研究背景 1
1.2 研究目的 2
1.3 主要貢獻 3
1.4 論文架構 3
第二章 相關研究 4
2.1 Local Binary Pattern (LBP) 4
2.2 Local Ternary Pattern (LTP) 5
2.3 Extended Local Ternary Pattern (ELTP) 6
2.4 Local Binary/Ternary Pattern的降維方式 7
2.4.1 合併直方圖相鄰元素 7
2.4.2 減少樣本數量 8
2.4.3 Center-Symmetric LBP 9
2.4.4 拆解圖形樣式 10
2.4.5 考慮旋轉不變的特性(rotational-invariance) 11
2.4.6 使用uniform pattern 11
2.5 Soft Histogram for Local Binary Pattern 14
2.6 Fuzzy C-means Clustering 16
2.7 Fuzzy K-medoids Clustering 17
第三章 延展式區域三元化圖型特徵描述子之比例式降維法 20
3.1 統計ELTP中uniform pattern於影像中出現比例 20
3.2 樣式間的距離定義 23
3.3 比例式降維法 25
3.4 樣式分群 29
3.4.1 DB(Davies Bouldin) index 29
3.4.2 Dunn index 30
3.4.3 以H2 distance計算樣式間距離 30
3.4.4 以K-medoids clustering對樣式分群 31
第四章 比例式降維法之實驗結果 32
4.1 紋理影像辨識 32
4.2 人臉辨識 38
4.3 抗噪性實驗 41
第五章 結合模糊理論與ELTP 45
5.1 模糊化的成員函式 46
5.2 利用GPU平行運算加速計算時間 46
5.3 在分群過程中加入模糊化機制 52
5.4 實驗結果 54
第六章 結論與未來展望 60
參考文獻 62
 
圖目錄

圖2.1：LBP基本定義 4
圖2.2：LTP(θ=5)受到輕微雜訊干擾並未影響其編碼 6
圖2.3：合併直方圖相鄰元素 8
圖2.4：LBP(8,1)與LBP(16,2) 9
圖2.5：LBP(4,1)、LBP(8,2)與LBP(12,3) 9
圖2.6：CS-LBP的計算方式 10
圖2.7：拆解LTP為兩組LBP 10
圖2.8：利用旋轉不變的特性降維 11
圖2.9：LBP(8,1)中的部分uniform pattern 12
圖2.10：UELTP1的部份樣式 12
圖2.11：UELTP2的部份樣式 13
圖2.12：UELTP3的部份樣式 13
圖2.13：UELTP4的部份樣式 13
圖2.14：成員函式f1,d(z)、f0,d(z) 15
圖3.1：ELTP(8,1)各類uniform pattern之集合示意圖 21
圖3.2：ELTP(8,1)各類uniform pattern在Brodatz紋理影像中出現的平均比例圖 22
圖3.3：VOC2012的部分測試影像 23
圖3.4：依出現比例將樣式分組的示意圖 28
圖4.1：紋理影像辨識實驗使用的部分樣本 32
圖4.2：影像受到雜訊干擾後之變化 33
圖4.3：各描述方式在影像受到不同雜訊干擾下的分類準確率 34
圖4.4：CDR-ELTP 在不同維度下的表現 35
圖4.5：CDR-ELTP 在不同樣式間距離下的表現 36
圖4.6：CDR-ELTP 中使用不同分群演算法的於紋理影像辨識的準確率比較圖 38
圖4.7：人臉辨識實驗中所使用的部分臉部影像 39
圖4.8：臉部區域劃分及權重分配示意圖 39
圖4.9：特徵串接示意圖 40
圖4.10：人臉辨識準確率比較圖 41
圖4.11：於影像中隨機取出的 patch 示意圖 42
圖4.12：影像受到雜訊影響後，特徵產生變化的平均次數 43
圖4.13：將最大距離正規化為1時，樣式編碼發生變化的平均距離 43
圖5.1：ELTP的編碼方式 45
圖5.2：模糊化的 ELTP 成員函式f0、f1與f2 46
圖5.3：Fuzzy ELTP之編碼方式說明 47
圖5.4：計算模糊化特徵時的一個thread block示意圖 49
圖5.5：計算模糊化特徵時的一個grid示意圖 50
圖5.6：計算soft histogram時的一個thread block示意圖 50
圖5.7：計算soft histogram時的一個grid示意圖 50
圖5.8：fuzzy ELTP在紋理影像實驗中的準確率比較圖 55
圖5.9：fuzzy ELTP中不同H值的效能比較圖 56
圖5.10：FKM-ELTP於紋理影像辨識準確率比較圖 57
圖5.11：fuzzy LBP與fuzzy ELTP抗噪性分析之比較圖 59
 
表目錄

表3.1：ELTP(8,1)各類uniform pattern於全部 6561 個樣式中佔有的比例 20
表3.2：ELTP(8,1)各類uniform pattern在Brodatz紋理影像中出現的平均比例 21
表3.3：ELTP(8,1)各類uniform pattern在VOC2012影像中出現的平均比例 23
表3.4：ELTP(8,1)以H2 distance定義之uniform pattern在6561個樣式中之比例 25
表3.5：重新定義後的各類uniform pattern在Brodatz紋理影像中出現的平均比例 25
表3.6：ELTP 以比例式降維法的舉例 28
表3.7：以兩種樣式間距離計算方式的ELTP(8,1)樣式分群結果衡量 31
表3.8：以兩種演算法對ELTP(8,1)樣式分群結果衡量 31
表4.1：各描述方式在影像受到不同雜訊干擾下的分類準確率 34
表4.2：CDR-ELTP在不同維度下的表現 35
表4.3：CDR-ELTP在不同樣式間距離下的表現 36
表4.4：CDR-ELTP中使用不同分群演算法於紋理影像辨識的準確率 37
表5.1：CPU與GPU計算96x96影像特徵的時間表 51
表5.2：CPU與GPU計算256x256影像特徵的時間表 52
表5.3：fuzzy ELTP 在紋理影像實驗中的準確率 54
表5.4：fuzzy ELTP中針對H值的實驗結果 56
表5.5：FKM-ELTP 於紋理影像辨識之準確率 57
表5.6：fuzzy LBP 與 fuzzy ELTP 抗噪性分析結果 59		zh_TW
dc.format.extent 6076967 bytes-
dc.format.mimetype application/pdf-
dc.language.iso en_US-
dc.source.uri (資料來源) http://thesis.lib.nccu.edu.tw/record/#G0098753030en_US
dc.subject (關鍵詞) 特徵描述zh_TW
dc.subject (關鍵詞) 區域二元化圖形zh_TW
dc.subject (關鍵詞) 延展式區域三元化圖形zh_TW
dc.subject (關鍵詞) 比例式降維法zh_TW
dc.subject (關鍵詞) 模糊理論zh_TW
dc.subject (關鍵詞) 紋理影像辨識zh_TW
dc.subject (關鍵詞) feature descriptoren_US
dc.subject (關鍵詞) local binary patternen_US
dc.subject (關鍵詞) extend local ternary patternen_US
dc.subject (關鍵詞) commensurate dimensionality reductionen_US
dc.subject (關鍵詞) fuzzy logicen_US
dc.subject (關鍵詞) texture classificationen_US
dc.title (題名) 延展式區域三元化圖形特徵描述子之比例式降維法zh_TW
dc.title (題名) Commensurate Dimensionality Reduction in Extended Local Ternary Patternsen_US
dc.type (資料類型) thesisen
dc.relation.reference (參考文獻) [1] T. Ojala, M. Pietikainen, and T. Maenpaa, “Multiresolution Gray-Scale and Rotation Invariant Texture Classification with Local Binary Patterns,” IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 24, no. 7, pp. 971-987, July 2002.
[2] D. Lowe, “Distinctive Image Features from Scale-Invariant Keypoints,” Int`l J. Computer Vision, vol. 2, no. 60, pp. 91-110, 2004.
[3] H. Bay, A. Ess, T. Tuytelaars, L. V. Gool, “SURF: Speeded Up Robust Features”, Computer Vision and Image Understanding (CVIU), Vol. 110, No. 3, pp. 346--359, 2008.
[4] A. Alahi, R. Ortiz, P. Vandergheynst, “FREAK: Fast Retina Keypoint”, IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2012.
[5] S. Leutenegger, M. Chli, R. Y. Siegwart, “BRISK: Binary Robust Invariant Scalable Keypoints”, IEEE International Conference on Computer Vision (ICCV), 2011.
[6] W. H. Liao, “Region Description Using Extended Local Ternary Patterns”, Proceedings of the 20th International Conference on Pattern Recognition, pp. 1003-1006, 2010.
[7] X. Tan and B. Triggs. “Enhanced local texture feature sets for face recognition under difficult lighting conditions”. In Analysis and Modeling of Faces and Gestures, volume 4778 of LNCS, pages 168–182. Springer, 2007.
[8] A. Shobeirinejad and Y. S. Gao, “Gender Classification Using Interlaced Derivative Patterns“, Proceedings of the 20th International Conference on Pattern Recognition, pp. 1509-1512, 2010.
[9] T. Ojala, M. Pietikainen and T. Maenpaa, ”Multi-resolution Gray-scale and Rotation Invariant Texture Classification with Local Binary Patterns”, IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 24(7), pp.971-987. July 2002.
[10] M. Heikkilä, M. Pietikäinen and C. Schmid, “Description of Interest Regions with Center-Symmetric Local Binary Patterns”, Computer Vision, Graphics and Image Processing, Lecture Notes in Computer Science, 2006.
[11] T. Ahonen, M. Pietikainen, “Soft Histograms for Local Binary Patterns”, in Proc. Fin. Signal Process. Symp., Oulu, Finland, 2007.
[12] J. C. Dunn, “A Fuzzy Relative of the ISODATA Process and Its Use in Detecting Compact Well-Separated Clusters”, Journal of Cybernetics 3: 32-57, 1973
[13] R. Krishnapuram, A. Joshi, L. Yi, “A Fuzzy Relative of the k-Medoids Algorithm with Application to Web Document and Snippet Clustering”, Snippet Clustering, in Proc. IEEE Intl. Conf. Fuzzy Systems - FUZZIEEE99, Korea, 1999.
[14] T. Ahonen, A. Hadid, and M. Pietikainen, “Face Description with Local Binary Patterns: Application to Face Recognition”, IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 28, no. 12, pp.2037-2041, Dec. 2006.
[15] N. P. Doshi, G. Schaefer, “A Comprehensive Benchmark of Local Binary Pattern Algorithms for Texture Retrieval”, International Conference on Pattern Recognition (ICPR), 2012.
[16] Everingham, M., Van~Gool, L., Williams, C. K. I., Winn, J., Zisserman, A., “The PASCAL Visual Object Classes Challenge 2012 (VOC2012) Results”, http://www.pascal-network.org/challenges/VOC/voc2012/workshop/index.html.		zh_TW