Symmetry is an important aspect that can be observed in various man made and natural objects. Symmetry is more likely to be a continuous feature rather than a binary feature. Quantitative analysis of the asymmetry, present in an object, is directly related to the measure of the difference between the given object and its symmetric counterpart. In our study, we propose a method, that reconstructs an image using Zernike moments, which represents the symmetric version of the given image. We use Zernike moments due to their invariance properties to various image transformations and ability to unique represent an image as the corresponding Zernike functions form an orthogonal basis. Ordinary least squares regression fails to calculate the Zernike moments. It is due to the presence of collinearity among the covariates obtained using Zernike functions. Therefore, we present the modified Ridge regression strategy to estimate Zernike moments, under a symmetry constraint, which enforces bilateral symmetry in planar images. An angle of the symmetry axis is estimated by minimizing the cost function, which determines the possible asymmetry present in the image. The given model is tested in three experiments involving image reconstruction and symmetry estimation. In the first experiment, images were reconstructed using Zernike moments calculated using Ridge regression. The second experiment consists of two parts. It starts with estimation of the symmetry axis evaluated for images showing bilateral symmetry at various angles. These experiments are then followed by image reconstruction with enforced symmetry. From these results the conclusion can be drawn that, Zernike moments estimated using Ridge regression can be efficiently used for the image reconstruction under the symmetry constraint. The symmetry constraint can be effectively utilized to estimate symmetry axis, and, in conjunction with Ridge regression, it can reconstruct an image which is the closest symmetric version of the given asymmetric image.