在腦科學計畫中，藉由研究果蠅的腦可能可以給我們更多關於人腦如何運作的資訊。在建立果蠅的腦的資料庫的時候，我們用共軛焦顯微鏡來擷取高解析度的影像。為了提高解析度，我們將一整個腦分成二或三或六個部份來拍攝影像，每一張影像都有與其它影像重疊的部份，而我們的首要工作就是要將這些影像拼接起來，以便利建立資料庫的進行。拼接的課題，不只在現在影像會遇到，這種提高解析度的拍攝方法在未來擷取更大實驗體的腦部影像時也必然會使用到，固有其重要性在。

　　為了拼接出共軛焦顯微鏡所拍攝出來的影像，我們做了以下這些努力：整個系統的架構分成兩個部份。第一個部份是找到影像之間相互拼接的相對位置。這個部份的問題包含了在同一平面去找尋影像之間的相對位置以及考慮在影像擷取中可能造成錯認為同一平面的誤差。第二個部份是將拼接的影像做亮度值的調整，將接縫盡量消除，減少人為拼貼的感覺。這個部份我們考慮到影像會因為每次重複取像而衰減的問題，所以使用了基於distance map, intensity compensating 以及multi-band blending的方法來做改進。最後，實驗的結果顯示這些方法的效果令人滿意而不需很多的參數設定。

By research on fruit flies’ brain circuits in life science, we could possibly get more information about how human brain works. One approach is to use high-resolution microscopy to take pictures of fruit flies’ brain slices of two or more images with overlap to each other. Combining these images is the first step to start the research on fruit flies’ brain circuits.
To reach the goal, we completed the work in two stages. In the first stage, the corresponding points in the two or more to-be-combined images are identified and registered. This part is composed of finding opposing positions of the images on x-y plane and taking adjacent images at z coordinate into consideration. In the second stage, the intensities of these images are blended. To eliminate the artifact of the seam as little as possible, distance map is computed. Besides, the intensity compensating method and the multi-band blending method are used. The experimental results show that the performance of the proposed method makes no requirement of setting many parameters and gives satisfactory results.

References

[1] Lowe, D. 2004. Distinctive image features from scale-invariant keypoints. International Journal of Computer Vision, 60(2):91–110.

[2] M. Brown and D. Lowe. Automatic Panoramic Image Stitching using Invariant Features. International Journal of Computer Vision, 2006

[3] G. M. Qu´enot. Computation of optical flow using dynamic programming. In Intl. Workshop on Machine Vision Applications, pages 249–52, 12-14 Nov. 1996.

[4] Alexei A. Efros and William T. Freeman. Image Quilting for Texture Synthesis and Transfer. Proceedings of SIGGRAPH 2001, August 2001

[5] Burt, P. and Adelson, E. 1983. A multiresolution spline with application to image mosaics. ACM Transactions on Graphics, 2(4):217–236.

[6] J. Davis. Mosaics of scenes with moving objects. In Proc. IEEE Conf. on Comp.Vision and Patt. Recog., 1998

[7] Y. Otha and T. Kanade, "Stereo by Intra- and Inter-Scanline Search Using Dynamic Programming" IEEE Trans. Pattern Analysis and Machine Intelligence, vol. PAMI-7, No.2, pp. 139-154, March 1985.

[8] B. Burg P. Missakian and B. Zavidovique, "Pattern Recognition through Dynamic Programming", SPIE's 29th Annual Internal Technical Symposium, San-Diego, July 1985.

[9] P. Adam, B. Burg and B. Zavidovique, "Dynamic Programming for Region Based Pattern Recognition", Proc. IEEE ICASSP-86, pp.2075-2078, April 1986.

[10] Jiaya Jia , Chi-Keung Tang, Eliminating Structure and Intensity Misalignment in Image Stitching, Proceedings of the Tenth IEEE International Conference on Computer Vision, p.1651-1658, October 17-20, 2005

[11] A. Zomet, A. Levin, S. Peleg, and Y. Weiss. Seamless image stitching by minimizing false edges. IEEE Trans. Image Processing, 2005.

[12] EDEN, A., UYTTENDAELE, M., AND SZELISKI, R. 2006. Seamless image stitching of scenes with large motions and exposure differences. Proceedings of CVPR 2006) 3, 2498–2505.

[13] Al-Kofahi, O., Can, A., Lasek, S., Szarowski, D.H., Turner, J.N. & Roysam, B. (2003) Algorithm for accurate 3-D registration of neuronal images acquired by confocal scanning laser microscopy. J. Microsc. 211, 000–000.

[14] Can, A., Al-Kofahi, O., Lasek, S., Szarowski, D. H., Turner, J. N. and Roysam, B. (2003). Attenuation correction in confocal laser microscopes: a novel two-view approach. J Microsc 211, 67-79.

[15] Hasler, D. and Susstrunk, S., 2004, Mapping colour in image stitching applications. Journal of Visual Communication and Image Representation, 15, 65-90.