【摘要】 本文简单介绍了最早的图像恢复的神经网络方法,Lambertian反射模型虽可用于图像的恢复与重建,但它有很大的局限性。为此,本文给出了一种基于改进的神经元网络的新型反射模型,用于三维图像的恢复与重建,此新法不需知道光源方向,经实验证明,具有计算快捷且图像识别精度高的特点。更多还原

【Abstract】 The paper introduces the earliest the image restoration based on the neural network. The limitation of the conventional reflectance model by Lambertian is also addressed, and a new neural - based reflectance model is proposed. The new neural learning algorithm is to optimize a proper reflectance model and to recover the object surface by a simple shape - form - shading(SFS) variational method with this improved neural - based reflectance model. An example is also given to prove that the SFS technique is robust for most objects, even when they are uncertain of h’ghting conditions.更多还原