Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2410.22139 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914996966916096 |
|---|---|
| author | Fu, Ruigang Hu, Qingyong Dong, Xiaohu Gao, Yinghui Li, Biao Zhong, Ping |
| author_facet | Fu, Ruigang Hu, Qingyong Dong, Xiaohu Gao, Yinghui Li, Biao Zhong, Ping |
| contents | As a fundamental operation in modern machine vision models, feature upsampling has been widely used and investigated in the literatures. An ideal upsampling operation should be lightweight, with low computational complexity. That is, it can not only improve the overall performance but also not affect the model complexity. Content-aware Reassembly of Features (CARAFE) is a well-designed learnable operation to achieve feature upsampling. Albeit encouraging performance achieved, this method requires generating large-scale kernels, which brings a mass of extra redundant parameters, and inherently has limited scalability. To this end, we propose a lightweight upsampling operation, termed Dynamic Lightweight Upsampling (DLU) in this paper. In particular, it first constructs a small-scale source kernel space, and then samples the large-scale kernels from the kernel space by introducing learnable guidance offsets, hence avoiding introducing a large collection of trainable parameters in upsampling. Experiments on several mainstream vision tasks show that our DLU achieves comparable and even better performance to the original CARAFE, but with much lower complexity, e.g., DLU requires 91% fewer parameters and at least 63% fewer FLOPs (Floating Point Operations) than CARAFE in the case of 16x upsampling, but outperforms the CARAFE by 0.3% mAP in object detection. Code is available at https://github.com/Fu0511/Dynamic-Lightweight-Upsampling. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2410_22139 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Lighten CARAFE: Dynamic Lightweight Upsampling with Guided Reassemble Kernels Fu, Ruigang Hu, Qingyong Dong, Xiaohu Gao, Yinghui Li, Biao Zhong, Ping Computer Vision and Pattern Recognition As a fundamental operation in modern machine vision models, feature upsampling has been widely used and investigated in the literatures. An ideal upsampling operation should be lightweight, with low computational complexity. That is, it can not only improve the overall performance but also not affect the model complexity. Content-aware Reassembly of Features (CARAFE) is a well-designed learnable operation to achieve feature upsampling. Albeit encouraging performance achieved, this method requires generating large-scale kernels, which brings a mass of extra redundant parameters, and inherently has limited scalability. To this end, we propose a lightweight upsampling operation, termed Dynamic Lightweight Upsampling (DLU) in this paper. In particular, it first constructs a small-scale source kernel space, and then samples the large-scale kernels from the kernel space by introducing learnable guidance offsets, hence avoiding introducing a large collection of trainable parameters in upsampling. Experiments on several mainstream vision tasks show that our DLU achieves comparable and even better performance to the original CARAFE, but with much lower complexity, e.g., DLU requires 91% fewer parameters and at least 63% fewer FLOPs (Floating Point Operations) than CARAFE in the case of 16x upsampling, but outperforms the CARAFE by 0.3% mAP in object detection. Code is available at https://github.com/Fu0511/Dynamic-Lightweight-Upsampling. |
| title | Lighten CARAFE: Dynamic Lightweight Upsampling with Guided Reassemble Kernels |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2410.22139 |