خوشه‌بندی تصاویر مقیاس بالا با مدل‌سازی نشانه‌های معنایی و شبیه‌سازی مکانی
محورهای موضوعی : مهندسی مخابرات
مهدی جلالی
1
(گروه برق، واحد نقده، دانشگاه آزاد اسلامی، نقده، ایران)
کلید واژه: استخراج ویژگی جدید, بافت, رنگ, تبدیل طیفی, بازیابی تصویر,
چکیده مقاله :
در سال‌های اخیر حاشیه‌نویسی تصاویر یکی از موضوعات تحقیقاتی فعال است. در این مقاله برای حاشیه‌نویسی تصاویر، تکنیک خوشه‌بندی تعاونی نیمه نظارت‌شده پیشنهاد می‌شود. روش‌های خوشه‌بندی به دلیل عدم نیاز به حاشیه‌نویسی بسیار موردتوجه هستند. برای دستیابی به بالاترین کارایی، نتایج خوشه‌بندی شش سیستم با فضای رنگ و معیار شباهت متفاوت با رأی اکثریت به‌صورت تعاونی، باهم ترکیب می‌شوند. در شرایطی که تعداد رأی‌ها برای یک تصویر کم باشد از بازخورد مرتبط برای حاشیه‌نویسی آن استفاده می‌شود. امروزه بیشتر از معیار شباهت خطی برای تعیین شباهت بین تصاویر استفاده می‌شود، ولی مدل‌های غیرخطی به دلیل نزدیکی به سیستم بینایی انسان می‌توانند کارایی بسیار بهتری داشته باشند، بدین منظور معیار شباهت غیرخطی KMRBF برای شبیه‌سازی بینایی انسان و بهبود نتایج بازیابی پیشنهاد می‌شود. آزمایش‌ها روی پایگاه داده تصاویر کورل و تصاویر ماهواره‌ای نشان می‌دهند که روش پیشنهادی دارای کارایی مناسبی است. با توجه به نتایج به‌دست‌آمده در پایگاه داده تصاویر ماهواره‌ای فضای رنگ YIQ دارای دقت بالاتری (به مقدار 5/82 درصد) است. همچنین سه فضای رنگ CIELab, HSV و YIQ دارای کارایی بالاتری هستند، چون در این فضاهای رنگی لومینانس از کرومینانس جدا بوده و این فضاهای رنگی به سیستم بینایی انسان نزدیک‌تر هستند.
In recent years, image annotation is one of the active research topics. In this article, a semi-supervised cooperative clustering technique is proposed for image annotation. Clustering methods are very popular because they do not require annotations. In order to achieve the highest efficiency, the clustering results of six systems with different color space and similarity criteria are cooperatively combined with the majority vote. When the number of votes for an image is low, relevant feedback is used to annotate it. One of the most important parts of the image retrieval system and clustering algorithm is determining the appropriate similarity criteria between images. Nowadays, the linear similarity criterion is mostly used to determine the similarity between images, but the nonlinear models can have much better performance due to their proximity to the human vision system, for this purpose, the KMRBF nonlinear similarity criterion is used to simulate vision. Humans and improvement of recovery results are suggested. Experiments on the Corel image database and satellite images show that the proposed method has good performance. According to the results obtained in the satellite image database, the YIQ color space has a higher accuracy (82.5%). Also, the three color spaces CIELab, HSV and YIQ have higher efficiency, because in these color spaces, luminance is separated from chrominance and these color spaces are closer to the human vision system.
[1] F. Cao , J. Liang and B. Liang, “ A new initialization method for categorical data clustering,” Expert Systems with Applications, vol. 36,no.7, pp. 10223–10228, 2009, doi:10.1016/j.eswa.2009.01.060.
[2] A. Amato and V. Lecce, “ A knowledge based approach for a fast image retrieval system,” Image and Vision Computing, vol. 26,no.11, pp. 1466–1480, 2008, doi:10.1016/j.imavis.2008.01.005.
[3] Y. Chen, J. Z. Wang and R. Krovetz , “ Cluster-Based Retrieval of Images by Unsupervised Learning,” in IEEE Transactions on Image Processing, vol. 14, no. 8, pp. 1187-1201, Aug. 2005, doi: 10.1109/TIP.2005.849770.
[4] R. Zhang and Zh.Zhang, “ BALAS:Empirical Bayesian learning in the relevance feedback for image retrieval,” Image and Vision Computing, vol. 24,no.3, pp.211–223, 2006, doi:10.1016/j.imavis.2005.11.004.
[5] T.h. Gevers and A.W.M. Smeulders, “ Image Search Engines An Overview,” University of Amsterdam 1098 SJ Amsterdam, The Netherlands, 2003.
[6] Y. Rui, T. S. Huang and S. Mehrotra, “Content-based image retrieval with relevance feedback in MARS, ” Proceedings of International Conference on Image Processing, 1997, pp. 815-818 vol.2, doi: 10.1109/ICIP.1997.638621.
[7] J.A. Hartigan and M.A .Wong, “Algorithm AS136: A k-means Clustering algorithm,” Applied Statistic, vol.28,no.1, pp. 100-108. 1979, doi:10.2307/2346830.
[8] C. Singh, and K.P. Kaur, “A fast and efficient image retrieval system based on color and texture features, ” Journal of Visual Communication and Image Representation,vol. 41, pp.225-238. 2016, doi:10.1016/j.jvcir.2016.10.002.
[9] J.M. Ho et al. “A novel content based image retrieval system using kmeans/knn with feature extraction, ” Computer Science and Information System, vol.9,no.4,pp.1645-1661,2012,doi:10.1109/ICSAI.2012.6223128.
[10] J. Mao and A.K. Jain, “Texture Classificatioon and Segmentation using Multi-Resolution Simultaneous Autoregressive Models, ” Pattern Recognition,vol.25,no.2,pp.173-188,2010, doi:10.1016/0031-3203(92)90099-5.
[11] M. Jalali and T. Sedghi, “Semi Supervised Feature Extraction for Filling Semantic Gap in Image Retrieval,” in Proc. of IEEE, Machine Vision and Image Processing Symposium, vol.1 no.2,pp. 345-350. 2011, doi:10.1109/IranianMVIP.2011.6121537.
[12] M. Jalali “Classification Percentage Enhancement of Segmentation Indexed Image based on Clustering Algorithm,” International journal of engineering & technology sciences ,vol.1, no. 1,pp. 1-4, 2014.
[13] T. N. Kipf and M. Welling, “Semi-supervised classification with graph convolutional networks,” Int. Conf. Learn. Represent. ICLR 2017 - Conf. Track Proc., pp. 1–14, 2017, doi:10.48550/arXiv.1609.02907.
[14] R. Zhang and Z. Zhang, “BALAS: Empirical Bayesian learning in the relevance feedback for image retrieval,” Image and Vision Computing ,vol.24, no.3,pp. 211–223,2006, doi:10.1016/j.imavis.2005.11.004.
[15] J. Smith, “Color for Image Retrieval”, Image Databases: Search and Retrieval of Digital Imagery, John Wiley & Sons, New York, pp.285-311,2001.
[16] A. Amato and V.D. Lecce, “A knowledge based approach for a fast image retrieval system,” Image and Vision Computing,vol. 26, no. 11,2008, doi:10.1016/j.imavis.2008.01.005.
[17] F. Hajiani, N. Parhizgar, and A. Keshavarz, “Hyperspectral Image Classification Using Low-Rank Representation and Spectral-Spatial Information, ” Journal of Communication Engineering, vol.11,no.43, pp.27-38. 2022. (in persian).
[18] M. Unser, “Texture classification and segmentation using wavelet frames, ” in IEEE Transactions on Image Processing, vol. 4, no. 11, pp. 1549-1560, Nov. 1995, doi: 10.1109/83.469936.
[19] K. P. Yip, D. W. Cheung and M. K. Ng, “On discovery of extremely low-dimensional clusters using semi-supervised projected clustering, ” 21st International Conference on Data Engineering (ICDE'05), 2005, pp. 329-340, doi: 10.1109/ICDE.2005.96.
[20] L. Nanni, A. Rigo, A. Lumini, and S. Brahnam, “Spectrogram classification using dissimilarity space,” Appl. Sci., vol. 10, no. 12, pp. 1–17, 2020, doi:10.3390/app10124176.
[21] Y. Chen, J. Z. Wang and R. Krovetz ,“CLUE: Cluster-Based Retrieval of Images by Unsupervised Learning ,” in IEEE Transactions on Image Processing, vol.14, no.8, pp. 1187 - 1201, 2005, doi: 10.1109/TIP.2005.849770.
[22] M.jalali, and T.Sedghi, “Extraction of Multiple Hybrid Features to Reduce the Semantic Vacuum with the Semi-Supervised Classification,” Journal of Communication Engineering., vol. 12,no.45, pp. 31-44, 2022(in persian).
[23] M. Flickner et al., “Query by image and video content: the QBIC system,” in Computer, vol. 28, no. 9, pp. 23-32, Sept. 1995, doi: 10.1109/2.410146.
[24] X. Zhang, M. Lei, D. Yang, Y. Wang and L. Ma, “Multi-scale curvature product for robust image corner detection in curvature scale space,” Pattern Recognition Letters, vol.28,no.5, 2007, doi:10.1016/j.patrec.2006.10.006.
[25] N. C. Shirazi, R. Hamzehyan, and A. Masoomi, “ The Comparison of Classification Algorithms for Remote Sensing Images,” Journal of Communication Engineering., vol. 5,no.17, pp. 31-38, 2015(in persian).
[26] A. Soltani-Farani, H. R. Rabiee, and S. A. Hosseini, “Spatial-aware dictionary learning for hyperspectral image classification, ” IEEE Trans. Geosci. Remote Sens., vol. 53,no.1, pp. 527-541, 2015, doi: 10.1109/TGRS.2014.2325067.
_||_[1] F. Cao , J. Liang and B. Liang, “ A new initialization method for categorical data clustering,” Expert Systems with Applications, vol. 36,no.7, pp. 10223–10228, 2009, doi:10.1016/j.eswa.2009.01.060.
[2] A. Amato and V. Lecce, “ A knowledge based approach for a fast image retrieval system,” Image and Vision Computing, vol. 26,no.11, pp. 1466–1480, 2008, doi:10.1016/j.imavis.2008.01.005.
[3] Y. Chen, J. Z. Wang and R. Krovetz , “ Cluster-Based Retrieval of Images by Unsupervised Learning,” in IEEE Transactions on Image Processing, vol. 14, no. 8, pp. 1187-1201, Aug. 2005, doi: 10.1109/TIP.2005.849770.
[4] R. Zhang and Zh.Zhang, “ BALAS:Empirical Bayesian learning in the relevance feedback for image retrieval,” Image and Vision Computing, vol. 24,no.3, pp.211–223, 2006, doi:10.1016/j.imavis.2005.11.004.
[5] T.h. Gevers and A.W.M. Smeulders, “ Image Search Engines An Overview,” University of Amsterdam 1098 SJ Amsterdam, The Netherlands, 2003.
[6] Y. Rui, T. S. Huang and S. Mehrotra, “Content-based image retrieval with relevance feedback in MARS, ” Proceedings of International Conference on Image Processing, 1997, pp. 815-818 vol.2, doi: 10.1109/ICIP.1997.638621.
[7] J.A. Hartigan and M.A .Wong, “Algorithm AS136: A k-means Clustering algorithm,” Applied Statistic, vol.28,no.1, pp. 100-108. 1979, doi:10.2307/2346830.
[8] C. Singh, and K.P. Kaur, “A fast and efficient image retrieval system based on color and texture features, ” Journal of Visual Communication and Image Representation,vol. 41, pp.225-238. 2016, doi:10.1016/j.jvcir.2016.10.002.
[9] J.M. Ho et al. “A novel content based image retrieval system using kmeans/knn with feature extraction, ” Computer Science and Information System, vol.9,no.4,pp.1645-1661,2012,doi:10.1109/ICSAI.2012.6223128.
[10] J. Mao and A.K. Jain, “Texture Classificatioon and Segmentation using Multi-Resolution Simultaneous Autoregressive Models, ” Pattern Recognition,vol.25,no.2,pp.173-188,2010, doi:10.1016/0031-3203(92)90099-5.
[11] M. Jalali and T. Sedghi, “Semi Supervised Feature Extraction for Filling Semantic Gap in Image Retrieval,” in Proc. of IEEE, Machine Vision and Image Processing Symposium, vol.1 no.2,pp. 345-350. 2011, doi:10.1109/IranianMVIP.2011.6121537.
[12] M. Jalali “Classification Percentage Enhancement of Segmentation Indexed Image based on Clustering Algorithm,” International journal of engineering & technology sciences ,vol.1, no. 1,pp. 1-4, 2014.
[13] T. N. Kipf and M. Welling, “Semi-supervised classification with graph convolutional networks,” Int. Conf. Learn. Represent. ICLR 2017 - Conf. Track Proc., pp. 1–14, 2017, doi:10.48550/arXiv.1609.02907.
[14] R. Zhang and Z. Zhang, “BALAS: Empirical Bayesian learning in the relevance feedback for image retrieval,” Image and Vision Computing ,vol.24, no.3,pp. 211–223,2006, doi:10.1016/j.imavis.2005.11.004.
[15] J. Smith, “Color for Image Retrieval”, Image Databases: Search and Retrieval of Digital Imagery, John Wiley & Sons, New York, pp.285-311,2001.
[16] A. Amato and V.D. Lecce, “A knowledge based approach for a fast image retrieval system,” Image and Vision Computing,vol. 26, no. 11,2008, doi:10.1016/j.imavis.2008.01.005.
[17] F. Hajiani, N. Parhizgar, and A. Keshavarz, “Hyperspectral Image Classification Using Low-Rank Representation and Spectral-Spatial Information, ” Journal of Communication Engineering, vol.11,no.43, pp.27-38. 2022. (in persian).
[18] M. Unser, “Texture classification and segmentation using wavelet frames, ” in IEEE Transactions on Image Processing, vol. 4, no. 11, pp. 1549-1560, Nov. 1995, doi: 10.1109/83.469936.
[19] K. P. Yip, D. W. Cheung and M. K. Ng, “On discovery of extremely low-dimensional clusters using semi-supervised projected clustering, ” 21st International Conference on Data Engineering (ICDE'05), 2005, pp. 329-340, doi: 10.1109/ICDE.2005.96.
[20] L. Nanni, A. Rigo, A. Lumini, and S. Brahnam, “Spectrogram classification using dissimilarity space,” Appl. Sci., vol. 10, no. 12, pp. 1–17, 2020, doi:10.3390/app10124176.
[21] Y. Chen, J. Z. Wang and R. Krovetz ,“CLUE: Cluster-Based Retrieval of Images by Unsupervised Learning ,” in IEEE Transactions on Image Processing, vol.14, no.8, pp. 1187 - 1201, 2005, doi: 10.1109/TIP.2005.849770.
[22] M.jalali, and T.Sedghi, “Extraction of Multiple Hybrid Features to Reduce the Semantic Vacuum with the Semi-Supervised Classification,” Journal of Communication Engineering., vol. 12,no.45, pp. 31-44, 2022(in persian).
[23] M. Flickner et al., “Query by image and video content: the QBIC system,” in Computer, vol. 28, no. 9, pp. 23-32, Sept. 1995, doi: 10.1109/2.410146.
[24] X. Zhang, M. Lei, D. Yang, Y. Wang and L. Ma, “Multi-scale curvature product for robust image corner detection in curvature scale space,” Pattern Recognition Letters, vol.28,no.5, 2007, doi:10.1016/j.patrec.2006.10.006.
[25] N. C. Shirazi, R. Hamzehyan, and A. Masoomi, “ The Comparison of Classification Algorithms for Remote Sensing Images,” Journal of Communication Engineering., vol. 5,no.17, pp. 31-38, 2015(in persian).
[26] A. Soltani-Farani, H. R. Rabiee, and S. A. Hosseini, “Spatial-aware dictionary learning for hyperspectral image classification, ” IEEE Trans. Geosci. Remote Sens., vol. 53,no.1, pp. 527-541, 2015, doi: 10.1109/TGRS.2014.2325067.