臺灣博碩士論文加值系統

在現今社會中，隨著人口結構以及生活型態的改變，飼養寵物的風氣越來越盛行，根據行政院農委會截至105年的統計資料顯示，在台灣的寵物犬數量大約有171萬隻。隨著飼養犬隻的比例提升，表示飼主的責任日益重要，儘管目前有著嚴謹的法律規定和罰金之規範，民眾棄養寵物的情況依然存在，而此問題也反映在流浪狗的議題上，它是造成流浪狗氾濫的主要原因之一，其餘則包括家犬走失、流浪狗的自然繁殖以及非法繁殖場的大量丟棄。當流浪狗的數量不斷增加時，會產生像是疾病傳染、公共衛生、咬人、交通意外等社會問題。
為了解決流浪狗所衍生出的社會問題，除了相關的法律規範之外，若要能有效地管理犬隻，就必須識別犬隻的身份。目前辨識犬隻的方式主要有兩種：寵物晶片以及寵物項圈，其各自皆有明顯的優缺點，像是前者的優點有：晶片的不易竊取性和提高對犬隻的追蹤性，而缺點則包含掃描的靈敏度、讀取器的相容性以及晶片對犬隻的健康疑慮；至於後者的優點有：免於犬隻承受植入晶片的疼痛，但卻有容易弄丟和被竊取的缺點存在。
因此本研究利用狗鼻紋紋路的唯一性，結合影像處理的技術，來建立一套狗鼻紋辨識系統，藉此解決上述方式所遇到的困難。狗鼻紋辨識系統主要分成兩個部分：狗鼻子影像切割及狗鼻紋影像辨識，本篇研究主要著墨於狗鼻紋影像辨識前的處理，其中包含狗鼻紋的紋路切割以及狗鼻紋影像的基準點定位兩大主軸。
本研究共使用61張狗鼻紋影像來進行實驗，分別來自24隻不同的犬隻，其實驗結果顯示整體的基準點定位準確率可以達81.98%。而在未來研究中，期望能改善本研究中所遇到的難題，並藉由同一犬隻不同影像間所尋找出的對應區塊來進行後續的犬隻辨識動作，讓此套狗鼻紋辨識系統的功能可以更加完整和完善。

With the change of the population structure and lifestyle, keeping pets is becoming more and more popular in today’s society. According to the statistics of the Council of Agriculture by 2016, the number of pet dogs in Taiwan is about 1.71 million. With the proportion of keeping dogs rise, the responsibility of the owners is more important. Although there are rigorous legal rules and fines at present, the situation of pet abandonment still exists. However, it is reflected in the problem of stray dogs, because it is one of the main reasons of the spread of stray dogs. The remaining reasons include the loss of the pet dogs and the natural propagation of stray dogs. When stray dogs increase continuously, it will lead to many social problems such as diseases, public health, bites, traffic accidents and so on. To solve the above-mentioned problems, the method of effective management for the dogs is identifying them. Currently, the main way to identify dogs is pet microchip and pet collar, but each of them has its own disadvantages. The shortcomings of the former include the sensitivity of the scan, reader’s compatibility and dog's health, while the latter is easy to lose or be stolen.
Because of these disadvantages, we want to develop a system which combines the unique dog nose prints and image processing. The system consists of two parts: nose image segmentation and nose prints recognition. In this study, we mainly focus on the processing before recognizing nose prints, including the segmentation of the dog nose prints and pivot point location of dog nose printimage. And in this paper, we use a total of 61 images that are respectively from 24 different dogs to conduct experiment. In experimental results, the accuracy of pivot point location is 81.98%. In the future work, it is expected to improve the difficulties encountering in this study, and use the different images belonging to the same dog to find out the corresponding block to identify each dog, so that the function of the dog nose pattern recognition system can be more completed and perfect.

摘要 i
Abstract ii
Table of Contents iv
List of Tables vi
List of Figures vii
Chapter 1 Introduction 1
Chapter 2 Related Works 5
2.1. Histogram Equalization 5
2.2. Gamma Correction 6
2.3. Run-Length 6
2.4. Multi-Scale Line Detector Method 8
2.5. Otsu’s Method 9
2.6. Mathematical Morphology 11
2.6.1. Dilation and Erosion 11
2.6.2. Closing 12
2.6.3. Thinning and Trimming Spurs 13
2.7. Region Labeling 14
Chapter 3 Segmentation of Dog Nose Pattern 18
3.1. Pre-processing 19
3.1.1. Convert the Original RGB Image to the Grayscale Image 20
3.1.2. Histogram Equalization 21
3.1.3. Localized Gamma Correction 21
3.1.4. Median Filter 23
3.1.5. Run-Length 24
3.2. Image Segmentation 24
3.2.1. Multi-Scale Line Detector Method 25
3.2.2. Hysteresis Thresholding 27
3.2.3. Closing 29
3.2.4. Thinning 29
3.2.5. Trimming Spurs 30
3.3. Adjustment after Image Segmentation 31
Chapter 4 Pivot Point Location of Dog Nose Image 33
4.1. Calculate the Center Point and the Nostrils’Distance 33
4.2. The Method of Region Search of Scale Resizing 34
4.3. Feature Matching Based on Genetic Algorithm 39
4.3.1. The Using Features 39
4.3.2. The Distance Algorithm 41
4.3.3. Genetic Algorithm Based Parameter Adjustment 42
Chapter 5 Experimental Results and Discussion 46
5.1. The Samples of Experiment 46
5.2. The Results of Experiment 47
5.2.1. Weight Values, Power Values and the Tolerance of Angle 47
5.2.2. The Label and Distance of the Blocks with Minimum Distance 48
5.2.3. The Accuracy of Proposed Method 49
5.3. The Challenges of Experiment and Discussion 51
5.3.1. The Angle of Shooting 51
5.3.2. The Quality of Shooting 51
5.3.3. The Error of Locating 52
Chapter 6 Conclusion and Future Work 54
References 56

[1] 行政院農委會畜牧處“10年內國人飼養毛小孩激增近百萬 飼主責任重要”
http://www.coa.gov.tw/theme_data.php?theme=news&sub_theme=agri&id=5503&print=Y

[2] J. McGrath. How Pet Microchipping Works. Available: http://science.howstuffworks.com/innovation/everyday-innovations/pet-microchip.htm

[3] L. T. Dickert. Dogs Noseprints Can be Used to Prove Identity, Just Like Fingerprints. Available:
http://www.allpetnews.com/dogs-noseprints-can-be-used-to-prove-identity-just-like-fingerprints

[4] N. Coldea. "Nose prints as a method of identification in dogs." Veterinary Quarterly 16.sup1 (1994): 60-60.

[5] K. S. Sim, C. P. Tso, and Y. Y. Tan. "Recursive sub-image histogram equalization applied to gray scale images." Pattern Recognition Letters 28.10 (2007): 1209-1221.

[6] P. Shanmugavadivu and K. Balasubramanian. "Particle swarm optimized multi-objective histogram equalization for image enhancement." Optics & Laser Technology 57 (2014): 243-251.

[7] X. Guan, S. Jian, P. Hongda, Z. Zhiguo, and G.Haibin. "An image enhancement method based on gamma correction." Computational Intelligence and Design, 2009. ISCID'09. Second International Symposium on. Vol. 1. IEEE (2009): 60-63.

[8] Z. Dongni, P. Won-Jae, L. Seung-Jun, A. C. Kang, and K. Sung-Jea. "Histogram partition based gamma correction for image contrast enhancement." Consumer Electronics (ISCE), IEEE 16th International Symposium on. IEEE (2012): 1-4.

[9] Y. K. Chan and C. C. Chang. "Image matching using run-length feature." Pattern Recognition Letters 22.5 (2001): 447-455.

[10] U. T. Nguyen, A. Bhuiyan, L. A. Park, K. Ramamohanarao. "An effective retinal blood vessel segmentation method using multi-scale line detection." Pattern recognition 46.3 (2013): 703-715.

[11] N. Otsu. "A threshold selection method from gray-level histograms." IEEE transactions on systems, man, and cybernetics 9.1 (1979): 62-66.

[12] R. M. Haralick, S. R. Sternberg, and X. Zhuang. "Image analysis using mathematical morphology." IEEE transactions on pattern analysis and machine intelligence 4 (1987): 532-550.

[13] J. Crespo and R. Schafer. Mathematical morphology and its applications to image processing. Vol. 2. Springer Science & Business Media (2012).

[14] Y. K. Chan, D. C. Huang, K. C. Liu, R. T. Chen, and X. Jiang. "An automatic indirect immunofluorescence cell segmentation system." Mathematical Problems in Engineering (2014).

[15] A. Rosenfeld. "Connectivity in digital pictures." Journal of the ACM (JACM) 17.1 (1970): 146-160.

[16] S. Levialdi. "On shrinking binary picture patterns." Communications of the ACM 15.1 (1972): 7-10.

[17] U. Maulik. "Medical image segmentation using genetic algorithms." IEEE Transactions on information technology in biomedicine 13.2 (2009): 166-173.

[18] M. Sezgin and B. Sankur. "Survey over image thresholding techniques and quantitative performance evaluation." Journal of Electronic imaging 13.1 (2004): 146-168.

[19] M. Vlachos and E. Dermatas "Multi-scale retinal vessel segmentation using line tracking." Computerized Medical Imaging and Graphics, 34 (3) (2010): 213-227.

[20] R. C. Gonzalez and R. E. Woods. Digital Image Processing, Addison-Wesley Longman Publishing Co., Inc., Boston, MA (2002).