Spatially Adaptive Image Restoration and its FIR Implementation

This paper presents a spatially adaptive image restoration for EDoF lens images. A practical implementation method is also discussed for real-time applications. For basic image restoration we adopt the truncated constrained least squares (TCLS) filter, and for adaptively selecting the weights of the observed and the restored images we compute local variance. Based on experimental results, we can claim that the proposed restoration algorithm can successfully remove the image degradation with suppressing noise amplification.