-
摘要: 随着指纹识别技术的广泛应用，大量指纹图像需要被收集和存储.在指纹识别系统中，对于大容量的指纹数据库，指纹图像必须经过压缩后存储以减少存储空间，本文提出了基于自适应稀疏变换的指纹图像压缩算法.该算法在离线状态下提取指纹图像特征训练超完备字典；在编码过程中，首先利用差分预测编码和稀疏变换将待压缩指纹图像转换到稀疏域，然后对直流系数和稀疏表达系数进行量化和熵编码，从而实现图像信息的压缩.实验表明，在中低码率段，本文算法相比于JPEG、JPEG2000和WSQ等主流压缩算法表现出更优越的率失真性能；在相同码率时，本文算法生成的压缩图像的主观视觉效果更好，指纹识别率更高.Abstract: With the wide application of fingerprint identification technology, a large number of fingerprint images need to be collected and stored. In fingerprint identification, as for the fingerprint database with large-capacity, the fingerprint images must be stored after compression to reduce the storage space. In this paper, a fingerprint image compression algorithm based on adaptive sparse transformation is proposed. The feature of the fingerprint image is extracted offline to train the over-complete dictionary. In the encoding process, the fingerprint image to be compressed is converted to sparse domain by utilizing the differential predictive coding and sparse transformation in the first place; after that the DC coefficients and the sparse coefficients are quantized and entropy coded to achieve the compression of the image information. Experimental results show that the proposed algorithm outperforms the mainstream compression methods, such as JPEG, JPEG2000 and WSQ, in terms of ratio-distortion performance of decoded fingerprint image, especially at low to medium bit rates. At the same bit rate, the compression image generated by the proposed algorithm exhibits better subjective visual effect and higher fingerprint recognition rate.
-
图 1 本文提出的基于自适应稀疏变换的指纹图像压缩算法框架
Fig. 1 The framework of the proposed fingerprint image compression algorithm via adaptive sparse transformation
图 2 本文算法低频预测图像(a)与K-SVD-SR 算法低频图像 (b)块效应对比
Fig. 2 The comparison of low-frequency predicted image between the proposed algorithm (a) and the K-SVD-SR algorithm(b)
图 3 块间像素预测对最终编解码效果的影响
Fig. 3 The effect of inter-block pixel prediction on the final codec
图 5 分块尺寸分别为6× 6、7× 7、8× 8、9×9的率失真性能比较
Fig. 5 The comparison of rate distortion performance between the blocks with the size of 6× 6,7× 7,8× 8,and 9× 9
图 6 (a)~(d) 分别表示测试图像库中的数据库2~5在四种压缩算法下的平均率失真性能
Fig. 6 The (a),(b),(c),(d) respectively denotes the average rate distortion performance of the test image library Database2,Database3,Database4,and Database5 at 4 compression algorithms
图 7 从左至右分别表示原始图像和码率同为0.1 bpp的JPEG、JPEG2000、K-SVD-SR和本文算法的解码图像
Fig. 7 From left to right respectively represents the original image and the decoded image of JPEG,JPEG2000,K-SVD-SR,and the proposed algorithm at the same rate as 0.1 bpp
图 8 稀疏度自适应选择与固定稀疏度L=2、6、10、14对比
Fig. 8 The contrast of the adaptive sparsity and the fixed sparsity of L=2,6,10,14
图 10 两种量化模式对图像压缩性能的影响比较
Fig. 10 The comparison of the impact on image compression performance between the two quantization modes
图 11 “索引-权值”编码模式与“原子个数-索引-权值”编码模式对比
Fig. 11 The contrast of the "index-weight" encoding mode and the "number of atoms-index-weight" encoding mode
图 12 基于AGR的图像解码与K-SVD-SR的 直接解码对比
Fig. 12 The contrast of the image decoding based on AGR and the direct decoding of K-SVD-SR
图 13 测试图像库数据库2在四种压缩算法下的\\图像的平均特征匹配率
Fig. 13 The average image feature matching rate of the test image library Database 2 at 4 compression algorithms
表 1 四种压缩算法的时间复杂度比较 (s)
Table 1 The comparison of time complexity about 4 compression algorithms (s)
码率 (bpp) 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 JPEG 0.16 0.16 0.16 0.16 0.17 0.17 0.17 0.17 0.17 JPEG2000 0.08 0.08 0.08 0.08 0.07 0.08 0.08 0.08 0.08 K-SVD-SR 1.26 1.74 2.20 2.64 3.19 3.61 4.13 4.49 4.89 本文算法 2.13 2.80 3.26 2.88 4.53 4.95 5.83 5.76 6.84 
下载: 导出CSV
-
[1] Pennebaker W B, Mitchell J L. JPEG:Still Image Data Compression Standard. US:Springer, 1993. http://cn.bing.com/academic/profile?id=1540900967&encoded=0&v=paper_preview&mkt=zh-cn [2] Marcellin M W, Gormish M J, Bilgin A, Boliek M P. An overview of JPEG-2000. In:Proceedings of the 2000 Data Compression Conference. Snowbird, UT:IEEE, 2000.523-541 [3] Bradley J N, Brislawn C M, Hopper T. FBI wavelet/scalar quantization standard for gray-scale fingerprint image compression. In:Proceedings of the SPIE 1961, Visual Information Processing Ⅱ. Orlando, FL:SPIE, 1993.293-304 [4] Skodras A, Christopoulos C, Ebrahimi T. The JPEG 2000 still image compression standard. IEEE Signal Processing Magazine, 2001, 18(5):36-58 doi: 10.1109/79.952804 [5] Shao G, Wu Y, Yong A, Liu X, Guo T. Fingerprint compression based on sparse representation. IEEE Transactions on Image Processing, 2014, 23(2):489-501 doi: 10.1109/TIP.2013.2287996 [6] Olshausen B A, Field D J. Sparse coding with an overcomplete basis set:a strategy employed by V1? Vision Research, 1997, 37(23):3311-3325 doi: 10.1016/S0042-6989(97)00169-7 [7] Emmanuel B, Mu'Azu M, Sani S, Garba S. A review of wavelet-based image processing methods for fingerprint compression in biometric application. British Journal of Mathematics and Computer Science, 2014, 4(19):2781-2798 doi: 10.9734/BJMCS [8] Qian C, Xu Z. Robust visual tracking via sparse representation under subclass discriminant constraint. IEEE Transactions on Circuits and Systems for Video Technology, 2016, 26(7):1293-1307 doi: 10.1109/TCSVT.2015.2424091 [9] Sun B, Liu Z, Sun Y, Su F, Cao L, Zhang H. Multiple objects tracking and identification based on sparse representation in surveillance video. In:Proceedings of the 2015 IEEE International Conference on Multimedia Big Data (BigMM). Beijing, China:IEEE, 2015.268-271 [10] Cheng M, Wang C, Li J. Single-image super-resolution in RGB space via group sparse representation. Iet Image Processing, 2015, 9(6):461-467 doi: 10.1049/iet-ipr.2014.0313 [11] Tropp J A. Greed is good:algorithmic results for sparse approximation. IEEE Transactions on Information Theory, 2004, 50(10):2231-2242 doi: 10.1109/TIT.2004.834793 [12] Mallat S G, Zhang Z F. Matching pursuits with time-frequency dictionaries. IEEE Transactions on Signal Processing, 1993, 41(12):3397-3415 doi: 10.1109/78.258082 [13] Pati Y C, Rezaiifar R, Krishnaprasad P S. Orthogonal matching pursuit:recursive function approximation with applications to wavelet decomposition. In:Proceedings of the 1993 Conference Record of the 27th Asilomar Conference on Signals, Systems, and Computers. Pacific Grove, CA:IEEE, 1993.40-44 http://cn.bing.com/academic/profile?id=185716565&encoded=0&v=paper_preview&mkt=zh-cn [14] Gharavi-Alkhansari M, Huang T S. A fast orthogonal matching pursuit algorithm. In:Proceedings of the 1998 IEEE International Conference on Acoustics, Speech, and Signal Processing. Seattle, WA:IEEE, 1998.1389-1392 http://dl.acm.org/citation.cfm?id=1892575 [15] Chen S S, Donoho D L, Saunders M A. Atomic decomposition by basis pursuit. SIAM Journal on Scientific Computing, 1998, 20(1):33-61 doi: 10.1137/S1064827596304010 [16] Zhu J Y, Wang Z Y, Zhong R, Qu S M. Dictionary based surveillance image compression. Journal of Visual Communication and Image Representation, 2015, 31:225-230 doi: 10.1016/j.jvcir.2015.07.002 [17] Setiawan A D, Suksmono A B, Mengko T L R, Gunawan H. Low-bitrate medical image compression. In:Proceedings of the ACA2011 IAPR Conference on Machine Vision Applications. Nara, Japan, 2011.544-547 [18] Xu J, Pi Y, Ming R. SAR image compression based on sparse representation. In:Proceedings of the 11th International Radar Symposium (IRS). Vilnius, Lithuania:IEEE, 2010.1-4 [19] Zhan X, Zhang R, Yin D, Huo C. SAR image compression using multiscale dictionary learning and sparse representation. IEEE Geoscience and Remote Sensing Letters, 2013, 10(5):1090-1094 doi: 10.1109/LGRS.2012.2230394 [20] Aharon M, Elad M, Bruckstein A. K-SVD:an algorithm for designing overcomplete dictionaries for sparse representation. IEEE Transactions on Signal Processing, 2006, 54(11):4311-4322 doi: 10.1109/TSP.2006.881199 [21] Anurakphanawan N, Lamsrichan P. Fingerprint recognition performance with WSQ, CAWDR, and JPEG2000 compression. In:Proceedings of the 6th International Conference of Information and Communication Technology for Embedded Systems (IC-ICTES). Hua-Hin, Thailand:IEEE, 2015.1-6 [22] Fingerprint images[Online], available:http://pan.baidu.com/s/1i3KxQZV,November27,2015 [23] Abraham J, Kwan P, Gao J B. Fingerprint matching using a hybrid shape and orientation descriptor. State of the Art in Biometrics. New York:InTech, 2011. -
计量
- 文章访问数: 2490
- HTML全文浏览量: 456
- PDF下载量: 746
- 被引次数: 0
