Search papers, labs, and topics across Lattice.
This paper introduces LogicIR, a novel image restoration model that utilizes logic gate networks (LGNs) to achieve efficient computation. By employing a UNet-inspired architecture composed entirely of logic gates and incorporating a differentiable bit decoding layer along with an index shuffling mechanism, LogicIR significantly enhances information propagation. Experimental results show that LogicIR not only maintains high restoration quality but also reduces computational costs, positioning it as a promising alternative in the field of image restoration.
LogicIR achieves high-quality image restoration with a fraction of the computational cost by leveraging the efficiency of logic gate networks.
Image restoration aims to reconstruct high-quality images from degraded low-quality inputs. As the computational demands of image restoration models continue to rise, there is growing interest in lightweight architectures optimized for fast and efficient inference. Logic gate networks (LGNs), which operate using fundamental logic operations such as NAND and XOR, have recently emerged as a promising direction for achieving highly efficient computation. However, their potential remains largely untapped in the domain of image restoration. In this work, we introduce LogicIR, the first LGN specifically designed for image restoration tasks. LogicIR incorporates a UNet-inspired architecture composed entirely of logic gates. In addition, we propose a differentiable bit decoding layer and an index shuffling mechanism that improves information propagation across logic gates. Experimental results across multiple image restoration benchmarks demonstrate that LogicIR achieves strong performance with significantly reduced computational cost, establishing LogicIR as a viable and efficient alternative for image restoration. The source code is available at https://github.com/jimmy9704/LogicIR
