Learning Point Spread Function Invertibility Assessment for Image Deconvolution

2024-05-25

European Signal Processing Conference

AI Summary

In blind image deconvolution, the absence of a quantifiable metric for point spread function (PSF) invertibility hinders both reconstruction quality improvement and the optimization design of diffractive optical elements (DOEs). Method: We propose a neural-network-based, nonlinear metric for PSF invertibility that maps the PSF to a unit impulse response; the differentiable, low-complexity mapping error serves as a surrogate for invertibility. Contribution/Results: This is the first end-to-end differentiable learning framework for PSF invertibility modeling. The metric exhibits strong correlation with both deep and classical deconvolution performance, significantly improves reconstruction quality, and reduces computational overhead substantially compared to condition-number-based methods. Furthermore, it is seamlessly integrated into an end-to-end DOE optimization pipeline, achieving state-of-the-art results across multiple benchmarks.

Technology Category

• Search and Optimization: Non-convex Optimization
• Computer Vision: Learning & Optimization for CV
• Machine Learning: Adversarial Learning & Robustness

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Security and Privacy: Large-scale security measurements
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web

Abstract

Deep-Iearning (DL)-based image deconvolution (ID) has exhibited remarkable recovery performance, surpassing traditional linear methods. However, unlike traditional ID approaches that rely on analytical properties of the point spread function (PSF) to achieve high recovery performance—such as specific spectrum properties or small conditional numbers in the convolution matrix—DL techniques lack quantifiable metrics for evaluating PSF suitability for DL-assisted recovery. Aiming to enhance deconvolution quality, we propose a metric that employs a non-linear approach to learn the invertibility of an arbitrary PSF using a neural network by mapping it to a unit impulse. A lower discrepancy between the mapped PSF and a unit impulse indicates a higher likelihood of successful inversion by a DL network. Our findings reveal that this metric correlates with high recovery performance in DL and traditional methods, thereby serving as an effective regularizer in deconvolution tasks. This approach reduces the computational complexity over conventional condition number assessments and is a differen-tiable process. These useful properties allow its application in designing diffractive optical elements through end-to-end (E2E) optimization, achieving invertible PSFs, and outperforming the E2E baseline framework.

Problem

• Point Spread Function
• Image Clarity Enhancement
• Optical Element Design

Innovation

• Neural Network Methodology
• Point Spread Function (PSF) Quality Assessment
• Optical Element Design Enhancement