Research map / Video / Video segmentation and tracking
Streaming Multiscale Deep Equilibrium Models
Can Ufuk Ertenli, Emre Akbas, Ramazan Gokberk Cinbis
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| warm-start the solver from the previous frame (StreamDEQ) with 4 iterations | MDEQ-XL · Cityscapes val | mIoU: 42.3 -> 71.5 | Improved |
| warm-start the solver from the previous frame (StreamDEQ) with 8 iterations | MDEQ-XL · Cityscapes val | mIoU: 73.2 -> 78.2 | Improved |
| warm-start from the previous frame with 1 iteration | MDEQ-XL · Cityscapes val | mIoU: 2.2 -> 44.9 | Improved |
| warm-start with 4 iterations | MDEQ-XL + Faster R-CNN · ImageNet-VID val | mAP@50: 8.2 -> 50.4 | Improved |
| warm-start with 8 iterations | MDEQ-XL + Faster R-CNN · ImageNet-VID val | mAP@50: 32.6 -> 54.8 | Improved |
| warm-start with 2 iterations | MDEQ-XL + Faster R-CNN · ImageNet-VID val | mAP@50: 0 -> 39.5 | Improved |
| use StreamDEQ 8 iterations instead of 26-iteration MDEQ | MDEQ-XL · ImageNet-VID val | mAP@50: 55.0 -> 54.8 | No clear effect |
| use StreamDEQ 8 iterations instead of 27-iteration MDEQ | MDEQ-XL · Cityscapes val | mIoU: 79.7 -> 78.2 | Hurt |
| initialise first frame with zeros instead of reference representation (ablation) | StreamDEQ · Cityscapes val | mIoU | No clear effect |
| start from a different-dataset frame (shot change) (ablation) | StreamDEQ · Cityscapes val | mIoU | No clear effect |
Closest papers by meaning
- Looking Backward: Streaming Video-to-Video Translation with Feature Banks (ICLR 2025 · similarity 0.85)
- Scaling Video-Language Models to 10K Frames via Hierarchical Differential Distillation (ICML 2025 · similarity 0.84)
- StreamForest: Efficient Online Video Understanding with Persistent Event Memory (NeurIPS 2025 · similarity 0.84)
- Streaming Long Video Understanding with Large Language Models (NeurIPS 2024 · similarity 0.84)
- Learning from Streaming Video with Orthogonal Gradients (CVPR 2025 · similarity 0.84)
- VideoLLM-MoD: Efficient Video-Language Streaming with Mixture-of-Depths Vision Computation (NeurIPS 2024 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
