针对经过饱和氯化钙染色的四肢周围神经MicroCT扫描图像，提出了一种自动分离神经束区域的方法.1）手工标注第1幅扫描图像中的神经束轮廓和非轮廓区域；2）采用稀疏自动编码机方法提取神经束轮廓和非轮廓的特征并训练神经网络，进而自动识别分离第2幅图像中神经束区域；3）将第2幅图像中的神经束轮廓作为标注样本去训练新的神经网络；4）不断循环该过程直至序列图像全部处理完成.实验结果表明，所提出的神经束轮廓自动分离方法可以达到手工分离精度的84.7%，处理完成522幅图像需要花费0.3~0.4 h.此外，该方法不仅能快速准确地分离出序列图像中的单根神经束，而且还能分离出分裂与合并阶段的神经束轮廓，具有比较强的适应性.
A novel algorithm to extract the region of peripheral nerve internal fascicular groups automatically from MicroCT images is proposed. 1) the contour and non-contour of fascicular groups in the first MicroCT image were marked manually. 2) the sparse auto encoder (SAE) was employed to extract the features unsupervisedly and the corresponding neural network was trained in order to recognize the contour in the second image. 3) the contour and non-contour of fascicular groups in the second MicroCT image were used as the labeled samples to train a new neural network. 4) above process was repeated until all images were processed completely. The experimental results show that the accuracy of the algorithm in this paper reaches 84.7% of manual's and it can process 522 images only in 0.3~0.4 h. Additionally, not only the single groups but also the splitting and emerging groups can be extracted rapidly and accurately.
[1] Kleoniki L, Nikolaos N, Ioannis P, Konstantinos P. Three computer methods to reconstruct pulpal blood vessels and nerves[J]. Journal of Endodontic, 1995, 21(10):501-504.
[2] Sun K, Hu P, Zhang F. Three-dimensional reconstruction and visualization of the median nerve from serial tissue sections[J]. Microsurgery, 2009, 29(7):573-577.
[3] Zhong Y C, Wang L P, Dong J, Zhang Y, Luo, P, Qi, J, Liu, X L, Xian C J. Threedimensional reconstruction of peripheral nerve internal fascicular groups[J]. Scientific Reports, 2015. DOI:10.1038/srep17168.
[4] John H, Francis G, Elizabeth O, Alan D, Za W L, Michael H. The neurovascular supply of the developing vertebral body:a microCT and histologic analysis of the basivertebral foramen, nerve, and vessels[J]. The Spine Journal, 2009, 9(10):146S.
[5] Watling C P, Lago N, Benmerah S, Fitzgerald J J, Tarte E. Novel use of X-ray microcomputed tomography to image rat sciatic nerve and integration into scaffold[J]. Journal of Neuroscience Methods, 2010, 188(1):39-44.
[6] Hopkins T M, Heilman A M, Liggett J A, Lasance K, Little K J. Combining microcomputed tomography with histology to analyze biomedical implants for peripheral nerve repair[J]. Journal of Neuroscience Methods, 2015. 255:122-130.
[7] Zhu S, Zhu Q T, Liu X L, Yang W H, Jian Y T, Zhou X, He B, Gu L Q, Yan L W, Lin T, Xiang J X, Qi J. Three-dimensional reconstruction of the microstructure of human acellular nerve allograft[J]. Scientific Reports, 2016. DOI:10.1038/srep30694.
[8] Yan L, Guo Y, Qi J, Zhu Q, Gu L, Zheng C, Lin T, Lu Y, Zeng Z, Yu S, Zhu S, Zhou X, Zhang X, Du Y, Yao Z, Lu Y, Liu X. Iodine and freeze-drying enhanced highresolution MicroCT imaging for reconstructing 3D intraneural topography of human peripheral nerve fascicles[J]. Journal of Neuroscience Methods, 2017, 287:58-67.
[9] 孙志军,薛磊,许阳明,王正.深度学习研究综述[J].计算机应用研究,2012, 29(8):2806-2810. Sun Z J, Xue L, Xu Y M, Wang Z. Overview of deep learning[J]. Application Research of Computers, 2012, 29(8):2806-2810. (in Chinese)
[10] Li S Z, Yu B, Wu W, Su S Z, Ji R R. Feature learning based on SAE-PCA network for human gesture recognition in RGBD images[J]. Neurocomputing, 2015, 151:565-573.
[11] Su S Z, Liu Z H, Xu S P, Li S Z, Ji R R. Sparse auto-encoder based feature learning for human body detection in depth image[J]. Signal Processing, 2015, 112:43-52.
[12] Shah S A A, Bennamoun M, Boussaid F. Iterative deep learning for image set based face and object recognition[J]. Neurocomputing, 2016, 174:866-874.
[13] Jiang J, Trundle P, Ren J. Medical image analysis with artificial neural networks[J]. Computer Medical Imaging Graph, 2010, 34(8):617-31.
[14] Cai X, Hou Y, Li C, Lee J H, Wee W G. Evaluation of two segmentation methods on MRI brain tissue structures[C]//200628th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, New York, 2006, 1-15:4822-4825.
