<p>This paper aims to explore the application of stochastic volatility models in the context of image processing, specifically for noise reduction in images using wavelet transforms. We develop a new two-dimensional stochastic volatility model and study its application, especially in image denoising. The one-dimensional stochastic volatility models are commonly used for modeling financial time series. Extending this one-dimensional stochastic volatility model into two dimensions can be used to statistically model the wavelet coefficients. This model can account for significant features of wavelet coefficients, including their non-stationarity, heavy-tailed marginal distribution, and the dependencies between the coefficients, and it provides a plausible model to estimate their variances. The volatilities are then estimated using the Kalman filtering technique and we employ the minimum mean square error estimator to estimate the clean wavelet image coefficients. Results are compared with other image-denoising techniques to demonstrate the efficacy of the proposed method.</p>