一种基于先验标记特征的精准图像配准算法

刘天弼; 冯瑞

doi:10.3969/j.issn.1000-5641.2021.03.008

一种基于先验标记特征的精准图像配准算法

doi: 10.3969/j.issn.1000-5641.2021.03.008

刘天弼,
冯瑞^,

复旦大学计算机科学技术学院, 上海　201203

基金项目: 国家重点研发计划（2017YFC0803702）

详细信息

作者简介:
刘天弼, 男, 博士研究生, 研究方向为计算机视觉及人工智能. E-mail: allenlew@163.com

通讯作者:
冯　瑞, 男, 教授, 博士生导师, 研究方向为计算机视觉、多媒体及模式识别. E-mail: fengrui@fudan.edu.cn

中图分类号: TP391.7
计量
- 文章访问数: 299
- HTML全文浏览量: 424
- PDF下载量: 38
- 被引次数: 0
出版历程
- 收稿日期: 2020-04-28
- 网络出版日期: 2020-06-12
- 刊出日期: 2021-05-01

An algorithm for precise image registration based on priori mark features

LIU Tianbi,
FENG Rui^,

School of Computer Science and Technology, Fudan University, Shanghai　201203, China

摘要

摘要: 基因测序仪在读取基因序列之前需要将镜头与基因芯片精准对齐, 提出了一种能够精确地计算视场（Field of View, FOV）与理想位置偏差的算法. 预先在基因芯片上的特定位置设置标记, 通过拍摄的图像分析视场与基因芯片的位置误差: 首先, 提取图像灰度特征捕捉标记位置以初步对齐视场中心位置; 其次, 捕捉标记上的多个关键点的坐标; 最后, 对关键点的坐标映射关系进行拟合, 即可计算出精确的坐标和角度偏差. 实践和实验分析表明, 使用设计的图像配准算法能够实现对视场与基因芯片间位置偏差计算的高精度估计.
- 图像配准 /
- 位置校准 /
- 基因芯片
Abstract: The use of a gene sequencer requires that the lens and gene chip are aligned accurately before base-calling. We propose an algorithm to calculate the deviation of the field of view (FOV) from the ideal position. Marks are set at locations on the gene chip in advance, so that the deviation in position of the lens relative to the gene chip can be analyzed. Firstly, the marked locations are captured by extracting grayscale features of the image to initially align the center of the FOV; secondly, the coordinates for multiple key points on the marks are captured; and finally, the location and angle deviations are calculated by mapping coordinates for the key points. Practical and experimental analysis show that the image registration algorithm designed in this paper can achieve a high-precision estimate for the position deviation between the FOV and the gene chip.
- image registration /
- position calibration /
- gene chip

HTML全文

图 1 3种track标记设计示意图

Fig. 1 Schematic diagram of three track mark designs

下载: 全尺寸图片幻灯片

图 2 于竖直方向截取图像做水平卷积示意图

Fig. 2 Schematic diagram of the image in the vertical direction for horizontal convolution

下载: 全尺寸图片幻灯片

图 3 卷积运算捕捉cross标记示意图

Fig. 3 Schematic diagram of the convolution operation to capture cross marks

下载: 全尺寸图片幻灯片

图 4 单个像素错位映射

Fig. 4 Mapping of single pixel misalignment

下载: 全尺寸图片幻灯片

图 5 卷积核长度与容错角度的关系

Fig. 5 Relationship between the length of the convolution kernel and the angle of fault tolerance

下载: 全尺寸图片幻灯片

图 6 不同的卷积核切面

Fig. 6 Cross sections of various convolution kernels

下载: 全尺寸图片幻灯片

图 7 通过卷积捕捉track标记

Fig. 7 Capturing a track mark by convolution

下载: 全尺寸图片幻灯片

图 8 不同卷积核在理想状态下的区分度

Fig. 8 Differentiation of various convolution kernels under ideal conditions

下载: 全尺寸图片幻灯片

图 9 不同卷积核在噪声环境下的区分度

Fig. 9 Differentiation of various convolution kernels in a noisy environment

下载: 全尺寸图片幻灯片

图 10 不同卷积核在串扰环境下的区分度

Fig. 10 Differentiation of various convolution kernels in a light crosstalk environment

下载: 全尺寸图片幻灯片

图 11 5种抗串扰卷积核切面

Fig. 11 Cross sections of five anti-crosstalk convolution kernels

下载: 全尺寸图片幻灯片

图 12 5种抗串扰效果

Fig. 12 Anti-crosstalk effects of five convolution kernels

下载: 全尺寸图片幻灯片

表 1 位置与角度配准误差

Tab. 1 Position and angle deviation in registration

仪器分辨率/像素	cell大小/像素	亮度串扰/%	track宽度/像素	track数量/个		位置误差/像素	角度误差/(°)
仪器分辨率/像素	cell大小/像素	亮度串扰/%	track宽度/像素	横向	竖向	位置误差/像素	角度误差/(°)
2 560 × 2 160	9 × 9	15	9	8	8	0.243 1	0.010 882
2 560 × 2 160	6 × 6	10	6	8	8	0.274 6	0.012 292
5 012 × 5 012	9 × 9	20	9	9	9	0.284 9	0.006 376
5 012 × 5 012	6 × 6	15	6	9	9	0.330 5	0.007 397

下载: 导出CSV

参考文献(25)

[1]	滕晓坤, 肖华胜. 基因芯片与高通量测序技术前景分析 [J]. 中国科学 C 辑: 生命科学, 2008, 38(10): 891-899.
[2]	赵晨晖. 基于相位相关的亚像素图像配准 [J]. 现代计算机, 2014, 20(3): 50-53.
[3]	BROWN L G. A survey of image registration techniques [J]. ACM Computing Surveys, 1992, 24(4): 325-376. doi: 10.1145/146370.146374
[4]	ZITOVA B, FLUSSER J. Image registration methods: A survey [J]. Image and Vision Computing, 2003, 21(11): 977-1000. doi: 10.1016/S0262-8856(03)00137-9
[5]	PLUIM J P W, MAINTZ J B A, VIERGEVER M A. Image registration by maximization of combined mutual information and gradient information [J]. IEEE Transactions on Medical Imaging, 2000, 19(8): 809-814. doi: 10.1109/42.876307
[6]	AMAURY D, MARCHAND E. Second-order optimization of mutual information for real-time image registration [J]. IEEE Transactions on Image Processing, 2012, 21(9): 4190-4203. doi: 10.1109/TIP.2012.2199124
[7]	闫小超, 魏生民, 汪焰恩, 等. 基于序贯相似度的AGV图像配准方法 [J]. 科学技术与工程, 2010, 10(3): 696-699. doi: 10.3969/j.issn.1671-1815.2010.03.023
[8]	WEI X X, STOCKER A A. Mutual information, fisher information, and efficient coding [J]. Neural Computation, 2016, 28(2): 305-326. doi: 10.1162/NECO_a_00804
[9]	LOWE D G. Distinctive image features from scale-invariant keypoints [J]. International Journal of Computer Vision, 2004, 60(2): 91-110. doi: 10.1023/B:VISI.0000029664.99615.94
[10]	BAY H, TUYTELAARS T, VAN GOOL L. SURF: Speeded up robust features [C]//European Conference on Computer Vision, Computer Vision – ECCV 2006. Berlin: Springer, 2006: 404-417. DOI: 10.1007/11744023_32.
[11]	SMITH S M, BRADY J M. SUSAN—A new approach to low level image processing [J]. International Journal of Computer Vision, 1997, 23(1): 45-78. doi: 10.1023/A:1007963824710
[12]	TRAJKOVIE M, HEDLEY M. Fast corner detection [J]. Image and Vision Computing, 1998, 16(2): 75-87. doi: 10.1016/S0262-8856(97)00056-5
[13]	LEUTENEGGER S, CHLI M, SIEGWART R Y. BRISK: Binary robust invariant scalable keypoints [C]//2011 International Conference on Computer Vision. IEEE, 2011: 2548-2555. DOI: 10.1109/ICCV.2011.6126542.
[14]	SHARK L K, KUREKIN A A, MATUSZEWSKI B J. Development and evaluation of fast branch-and-bound algorithm for feature matching based on line segments [J]. Pattern Recognition, 2007, 40(5): 1432-1450. doi: 10.1016/j.patcog.2006.10.022
[15]	DAI X L, KHORRAM S. Development of a feature-based approach to automated image registration for multitemporal and multisensor remotely sensed imagery [C]//1997 IEEE International Geoscience and Remote Sensing Symposium Proceedings. Remote Sensing - A Scientific Vision for Sustainable Development. IEEE, 1997: 243-245.DOI: 10.1109/IGARSS.1997.61585.
[16]	GOVINDU V, SHEKHAR C, CHELLAPPA R. Using geometric properties for correspondence-less image alignment [C]//Proceedings of the 14th International Conference on Pattern Recognition (Cat. No.98EX170). IEEE, 1998: 37-41. DOI: 10.1109/ICPR.1998.711074.
[17]	HOLM M. Toward automatic rectification of satellite images using feature based matching [C]//Proceedings of the IGARSS’91 Remote Sensing: Global Monitoring for Earth Management. IEEE, 1991: 2439-2442. DOI: 10.1109/IGARSS.1991.575537.
[18]	HSIEH Y C, MCKEOWN D M, PERLANT F P. Performance evaluation of scene registration and stereo matching for cartographic feature extraction [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1992, 14(2): 214-238. doi: 10.1109/34.121790
[19]	LITJENS G, KOOI T, BEJNORDI B E, et al. A survey on deep learning in medical image analysis [J]. Medical Image Analysis, 2017, 2017, 42: 60-88. doi: 10.1016/j.media.2017.07.005
[20]	SOKOOTI H, DE VOS B, BERENDSEN F, et al. Nonrigid image registration using multi-scale 3D convolutional neural networks [C]//Medical Image Computing and Computer Assisted Intervention Society, MICCAI(2017), Part I, LNCS 10433. Berlin: Springer, 2017: 232-239. DOI: 10.1007/978-3-319-66182-7_27.
[21]	CAO X H, YANG J H, ZHANG J, et al. Deformable image registration using a cue-aware deep regression network [J]. IEEE Transactions on Biomedical Engineering, 2018, 65(9): 1900-1911. doi: 10.1109/TBME.2018.2822826
[22]	MIAO S, WANG Z J, LIAO R. A CNN regression approach for real-time 2D/3D registration [J]. IEEE Transactions on Medical Imaging, 2016, 35(5): 1352-1363. doi: 10.1109/TMI.2016.2521800
[23]	DALCA A V, BALAKRISHNAN G, GUTTAG J, et al. Unsupervised learning for fast probabilistic diffeomorphic registration[C]//International Conference on Medical Image Computing and Computer-Assisted Intervention–MICCAI 2018: Medical Image Computing and Computer Assisted Intervention. Cham: Springer, 2018: 729-738. DOI: 10.1007/978-3-030-00928-1_82.
[24]	DE VOS B D, BERENDSEN F F, VIERGEVER M A, et al. End-to-end unsupervised deformable image registration with a convolutional neural network [C]//International Workshop on Deep Learning in Medical Image Analysis–2017, International Workshop on Multimodal Learning for Clinical Decision Suppor–2017: Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support. Cham: Springer, 2017: 204-212. DOI: 10.1007/978-3-319-67558-9_24.
[25]	KREBS J, MANSI T, MAILHÉ B, et al. Learning structured deformations using diffeomorphic registration [EB/OL]. (2018-07-20)[2020-02-28]. https://arxiv.org/pdf/1804.07172v2.pdf.