Résumé
Objective: To compare the applications of five kinds of routine elastic registration methods included of Horn-Schunck optical flow method, Demons algorithm, accelerated Demons algorithm, level set algorithm and fast free deformation algorithm in radiotherapy. Methods: 10 patients who underwent radiotherapy on pelvic cavity were enrolled in the research. And their CT images at the time of location and during radiotherapy were collected, and the five kinds of routine elastic registration methods were applied in the research, and through autonomic programming to implement elastic registration, and then the registered results were further compared. Results: The average minimum mean square error (MSE) of five kinds of routine elastic registration methods have been decreased 6.7%-26.0%, and the average correlation coefficient(CC) has been increased 2.6%-3.7%, and the average registration time was 107.5s-220.8s. Conclusions: All of the five kinds of elastic registration methods get better results, and it is recommended to use Horn-Schunck optical flow method for elastic registration of radiotherapy CT. And through using elastic registration model can more accurate simulate the movement of organ, and it is a direction of further study that monitor individual dosage of organ at risk.
Résumé
Objective: To apply level set motion-based elastic registration method to radiotherapy CT image, then to provide technical support for the accurate evaluation of tumor and changing process of patients with endanger organ and its accumulative dose. Methods: Based on Vemuri’s level set motion method, we wrote the algorithm program using Matlab software and applied on two sets of 3D CT images from patients with cervical cancer and NPC for fully automatic elastic registration. Results: Comparing CT images before and after registration, for the cervical cancer patient, the minimum mean square error (MSE) decreased by 55.1%and correlation coefficient (CC) increased by 5.3%. For the NPC patient, MSE decreased by 32.1%and CC increased by 4.6%. Conclusion: From the image difference and evaluation parameters, the efficacy of level set motion-based elastic registration method was preliminarily demonstrated. In order to apply this method to clinical radiotherapy, dit needs to find a more accurate mathematical algorithm further in order to compute human anatomy deformation through image motion.
Résumé
En este artículo se propone un enfoque no paramétrico para el registro elástico de imágenes médicas multimodales, cuya idea principal radica en el uso de medidas de variabilidad local, basadas en la entropía, la varianza o una combination de ambas. La metodología empleada consiste en encontrar el campo vectorial de los desplazamientos entre los pixeles de las imágenes candidata y patrón empleando una tecnica compuesta por tres pasos: primero, se obtiene una aproximación del campo vectorial por medio de un registro paramétrico entre ambas imágenes; segundo, se mapean las imágenes registradas paramétricamente a un espacio de intensidades donde pueden ser comparadas; tercero, se obtiene el flujo óptico entre las imágenes en el espacio al que fueron mapeadas. El algoritmo propuesto se evalúo usando un conjunto de imágenes de resonancia magnética y tomografía computarizada adquiridas desde diferentes vistas, las cuales fueron deformadas sintéticamente. Los resultados obtenidos en la estimación del campo de desplazamientos con las cuatro medidas de variabilidad local propuestas muestran un error medio menor que 1.4 mm, y en el caso de la entropía menor a 1 mm. Además, se demuestra la convergencia del algoritmo con ayuda de la entropía conjunta. Asó, la metodología descrita representa una nueva alternativa para el registro elástico multimodal de imágenes médicas.
In this work, we present a novel approach for multimodal elastic registration of medical images, where the key idea is to use local variability measures based on entropy, variance or a combination of these metrics. The proposed methodology relies on finding the displacements vector field between pixels of a source image and a target one, using the following three steps: first, an initial approximation of the vector field is achieved by using a parametric registration based on particle filtering between the images to align; second, the images previously registered are mapped to a common space where their intensities can be compared; and third, we obtain the optical flow between the images in this new space. To evaluate the proposed algorithm, a set of computed tomography and magnetic resonance images obtained in different views, were modified with synthetic deformation fields. The results obtained with the four proposed local variability measures show an average error of less than 1.4 mm, and in the case of the entropy less than 1 mm. In addition, the convergence of the algorithm is highlighted by the joint entropy. Therefore, the described methodology could be considered as a new alternative for multimodal elastic registration of medical images.
Résumé
Objective To present a new algorithm for multidimensional medical image registration from global registration to local registration in sequence. Methods Firstly, the global registration was achieved by the method of affine transformation composed of B-splines,whose knots were the four vertexes of the medical image. Then the knots of the B-splines were increased, and the transformation function was more complex and elastic than ever,which completed the elastic aligning for the detail of the medical image. Results The whole registration algorithm represented the principle aligning from global registration to local registration. Conclusion It is proved by experiments that the presented algorithm can decrease the time of calculation and increase the robustness of registration.