Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
97 tokens/sec
GPT-4o
53 tokens/sec
Gemini 2.5 Pro Pro
44 tokens/sec
o3 Pro
5 tokens/sec
GPT-4.1 Pro
47 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Improving Unsupervised Video Object Segmentation via Fake Flow Generation (2407.11714v1)

Published 16 Jul 2024 in cs.CV

Abstract: Unsupervised video object segmentation (VOS), also known as video salient object detection, aims to detect the most prominent object in a video at the pixel level. Recently, two-stream approaches that leverage both RGB images and optical flow maps have gained significant attention. However, the limited amount of training data remains a substantial challenge. In this study, we propose a novel data generation method that simulates fake optical flows from single images, thereby creating large-scale training data for stable network learning. Inspired by the observation that optical flow maps are highly dependent on depth maps, we generate fake optical flows by refining and augmenting the estimated depth maps of each image. By incorporating our simulated image-flow pairs, we achieve new state-of-the-art performance on all public benchmark datasets without relying on complex modules. We believe that our data generation method represents a potential breakthrough for future VOS research.

User Edit Pencil Streamline Icon: https://streamlinehq.com
Authors (7)
  1. Suhwan Cho (33 papers)
  2. Minhyeok Lee (47 papers)
  3. Jungho Lee (25 papers)
  4. Donghyeong Kim (11 papers)
  5. Seunghoon Lee (28 papers)
  6. Sungmin Woo (16 papers)
  7. Sangyoun Lee (67 papers)

Summary

We haven't generated a summary for this paper yet.