Physics-based solver reconstructs images from fewer measurements faster
Computational compressive imaging reconstruction with a simulated-annealing-based Ising solver
Computer Vision and Pattern RecognitionDistributed, Parallel, and Cluster Computing
Summary
Capturing detailed images usually requires lots of data and strong computers, which can be slow and expensive. This paper shows how problems in creating images from limited data can be transformed into a physics puzzle that naturally finds the best picture. The authors use a method inspired by how physical systems cool down to solve these puzzles and recover high-quality images with less computation. Their approach works well for different imaging systems and could pave the way for faster, energy-efficient imaging tools.
What this means in practice
- •For camera hardware engineers: Build cameras that reconstruct images with less computing power by embedding physics-inspired solvers that handle inverse problems directly.
- •For medical imaging device teams: Implement fast image reconstruction algorithms for low-power imaging devices, enhancing portable and real-time medical diagnostics.
Authors
Ze Zheng, Xiaoyan Wu, Hang Xu, Jingzheng Huang, Tailong Xiao, Guihua Zeng
Abstract
Computational compressive imaging, such as single-pixel imaging and lensless imaging, is a promising imaging technique, enabling the reconstruction of objects from under-sampled measurements through compressed sensing algorithms or deep learning methods. However, the demand for electronic computational resources grows rapidly when recovering high-resolution and multi-dimensional images and training neural networks with numerous parameters, which has become a bottleneck for the widespread adoption of this technique. Here, we propose a physics-driven reconstruction framework that formulates imaging inverse problems as Ising models and directly recovers target objects through the self-evolution of non-von Neumann Ising machines toward the ground state of the Hamiltonian. We employ a simulated-annealing-based Ising solver to demonstrate this proof-of-concept by reconstructing objects in both single-pixel imaging and a lensless diffuser camera at sampling rates ranging from 1% to 25%. By further incorporating a foreground-sparsity constraint and a smoothness constraint, the reconstruction quality of our method is comparable to or even better than that of mainstream algorithms. This approach has the potential to substantially reduce the electronic computational burden of inverse reconstruction, paving the way for real-time, low-power computational imaging.