Saved in:
Bibliographic Details
Main Authors: Choi, Haram, Na, Cheolwoong, Oh, Jihyeon, Lee, Seungjae, Kim, Jinseop, Choe, Subeen, Lee, Jeongmin, Kim, Taehoon, Yang, Jihoon
Format: Preprint
Published: 2023
Subjects:
Online Access:https://arxiv.org/abs/2305.11474
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929319099498496
author Choi, Haram
Na, Cheolwoong
Oh, Jihyeon
Lee, Seungjae
Kim, Jinseop
Choe, Subeen
Lee, Jeongmin
Kim, Taehoon
Yang, Jihoon
author_facet Choi, Haram
Na, Cheolwoong
Oh, Jihyeon
Lee, Seungjae
Kim, Jinseop
Choe, Subeen
Lee, Jeongmin
Kim, Taehoon
Yang, Jihoon
contents Although many recent works have made advancements in the image restoration (IR) field, they often suffer from an excessive number of parameters. Another issue is that most Transformer-based IR methods focus only on either local or global features, leading to limited receptive fields or deficient parameter issues. To address these problems, we propose a lightweight IR network, Reciprocal Attention Mixing Transformer (RAMiT). It employs our proposed dimensional reciprocal attention mixing Transformer (D-RAMiT) blocks, which compute bi-dimensional (spatial and channel) self-attentions in parallel with different numbers of multi-heads. The bi-dimensional attentions help each other to complement their counterpart's drawbacks and are then mixed. Additionally, we introduce a hierarchical reciprocal attention mixing (H-RAMi) layer that compensates for pixel-level information losses and utilizes semantic information while maintaining an efficient hierarchical structure. Furthermore, we revisit and modify MobileNet V1 and V2 to attach efficient convolutions to our proposed components. The experimental results demonstrate that RAMiT achieves state-of-the-art performance on multiple lightweight IR tasks, including super-resolution, color denoising, grayscale denoising, low-light enhancement, and deraining. Codes are available at https://github.com/rami0205/RAMiT.
format Preprint
id arxiv_https___arxiv_org_abs_2305_11474
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Reciprocal Attention Mixing Transformer for Lightweight Image Restoration
Choi, Haram
Na, Cheolwoong
Oh, Jihyeon
Lee, Seungjae
Kim, Jinseop
Choe, Subeen
Lee, Jeongmin
Kim, Taehoon
Yang, Jihoon
Computer Vision and Pattern Recognition
Artificial Intelligence
Although many recent works have made advancements in the image restoration (IR) field, they often suffer from an excessive number of parameters. Another issue is that most Transformer-based IR methods focus only on either local or global features, leading to limited receptive fields or deficient parameter issues. To address these problems, we propose a lightweight IR network, Reciprocal Attention Mixing Transformer (RAMiT). It employs our proposed dimensional reciprocal attention mixing Transformer (D-RAMiT) blocks, which compute bi-dimensional (spatial and channel) self-attentions in parallel with different numbers of multi-heads. The bi-dimensional attentions help each other to complement their counterpart's drawbacks and are then mixed. Additionally, we introduce a hierarchical reciprocal attention mixing (H-RAMi) layer that compensates for pixel-level information losses and utilizes semantic information while maintaining an efficient hierarchical structure. Furthermore, we revisit and modify MobileNet V1 and V2 to attach efficient convolutions to our proposed components. The experimental results demonstrate that RAMiT achieves state-of-the-art performance on multiple lightweight IR tasks, including super-resolution, color denoising, grayscale denoising, low-light enhancement, and deraining. Codes are available at https://github.com/rami0205/RAMiT.
title Reciprocal Attention Mixing Transformer for Lightweight Image Restoration
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2305.11474