Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2404.08135 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911837375692800 |
|---|---|
| author | Lin, Jamie Menjay Jeong, Jisoo Cai, Hong Garrepalli, Risheek Wang, Kai Porikli, Fatih |
| author_facet | Lin, Jamie Menjay Jeong, Jisoo Cai, Hong Garrepalli, Risheek Wang, Kai Porikli, Fatih |
| contents | Optical flow estimation is crucial to a variety of vision tasks. Despite substantial recent advancements, achieving real-time on-device optical flow estimation remains a complex challenge. First, an optical flow model must be sufficiently lightweight to meet computation and memory constraints to ensure real-time performance on devices. Second, the necessity for real-time on-device operation imposes constraints that weaken the model's capacity to adequately handle ambiguities in flow estimation, thereby intensifying the difficulty of preserving flow accuracy. This paper introduces two synergistic techniques, Self-Cleaning Iteration (SCI) and Regression Focal Loss (RFL), designed to enhance the capabilities of optical flow models, with a focus on addressing optical flow regression ambiguities. These techniques prove particularly effective in mitigating error propagation, a prevalent issue in optical flow models that employ iterative refinement. Notably, these techniques add negligible to zero overhead in model parameters and inference latency, thereby preserving real-time on-device efficiency. The effectiveness of our proposed SCI and RFL techniques, collectively referred to as SciFlow for brevity, is demonstrated across two distinct lightweight optical flow model architectures in our experiments. Remarkably, SciFlow enables substantial reduction in error metrics (EPE and Fl-all) over the baseline models by up to 6.3% and 10.5% for in-domain scenarios and by up to 6.2% and 13.5% for cross-domain scenarios on the Sintel and KITTI 2015 datasets, respectively. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2404_08135 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | SciFlow: Empowering Lightweight Optical Flow Models with Self-Cleaning Iterations Lin, Jamie Menjay Jeong, Jisoo Cai, Hong Garrepalli, Risheek Wang, Kai Porikli, Fatih Computer Vision and Pattern Recognition Optical flow estimation is crucial to a variety of vision tasks. Despite substantial recent advancements, achieving real-time on-device optical flow estimation remains a complex challenge. First, an optical flow model must be sufficiently lightweight to meet computation and memory constraints to ensure real-time performance on devices. Second, the necessity for real-time on-device operation imposes constraints that weaken the model's capacity to adequately handle ambiguities in flow estimation, thereby intensifying the difficulty of preserving flow accuracy. This paper introduces two synergistic techniques, Self-Cleaning Iteration (SCI) and Regression Focal Loss (RFL), designed to enhance the capabilities of optical flow models, with a focus on addressing optical flow regression ambiguities. These techniques prove particularly effective in mitigating error propagation, a prevalent issue in optical flow models that employ iterative refinement. Notably, these techniques add negligible to zero overhead in model parameters and inference latency, thereby preserving real-time on-device efficiency. The effectiveness of our proposed SCI and RFL techniques, collectively referred to as SciFlow for brevity, is demonstrated across two distinct lightweight optical flow model architectures in our experiments. Remarkably, SciFlow enables substantial reduction in error metrics (EPE and Fl-all) over the baseline models by up to 6.3% and 10.5% for in-domain scenarios and by up to 6.2% and 13.5% for cross-domain scenarios on the Sintel and KITTI 2015 datasets, respectively. |
| title | SciFlow: Empowering Lightweight Optical Flow Models with Self-Cleaning Iterations |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2404.08135 |