Research map / Video / Video segmentation and tracking
Predicting Future Frames Using Retrospective Cycle GAN
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| use Retrospective Cycle GAN instead of ContextVP | video prediction on Caltech pedestrian · Caltech pedestrian | MSE: 1.94 -> 1.61 | Helped |
| use Retrospective Cycle GAN instead of ContextVP | video prediction on Caltech pedestrian · Caltech pedestrian | SSIM: 0.921 -> 0.919 | Hurt |
| use Retrospective Cycle GAN instead of ContextVP | video prediction on UCF101 · UCF101 | PSNR: 34.9 -> 35.0 | Helped |
| use Retrospective Cycle GAN instead of Liu et al. | video prediction on CUHK Avenue · CUHK Avenue | MSE: 0.51 -> 0.39 | Helped |
| use Retrospective Cycle GAN instead of Liu et al. | video prediction on ShanghaiTech · ShanghaiTech | PSNR: 31.4 -> 34.1 | Helped |
| use Retrospective Cycle GAN instead of PredNet | multi-step video prediction on Caltech pedestrian · Caltech pedestrian, T=15 | PSNR: 17.5 -> 18.4 | Helped |
| add retrospective loss to forward prediction (ablation) | Retrospective Cycle GAN · Caltech pedestrian | PSNR: 26.3 -> 26.8 | Helped |
| add backward prediction (ablation) | Retrospective Cycle GAN · Caltech pedestrian | PSNR: 26.8 -> 26.9 | Helped |
| add LoG reconstruction loss (ablation) | bidirectional Retrospective Cycle GAN · Caltech pedestrian | PSNR: 26.9 -> 27.5 | Helped |
| add frame discriminator (ablation) | bidirectional Retrospective Cycle GAN · Caltech pedestrian | PSNR: 27.5 -> 28.4 | Helped |
| add sequence discriminator (ablation) | Retrospective Cycle GAN with frame discriminator · Caltech pedestrian | PSNR: 28.4 -> 29.2 | Helped |
| vary input sequence length from 2 to 10 images (ablation) | Retrospective Cycle GAN · Caltech pedestrian | PSNR: 29.167; 29.222; 29.006; 28.940; 29.009 | No clear effect |
Closest papers by meaning
- Temporal Coherency based Criteria for Predicting Video Frames using Deep Multi-stage Generative Adversarial Networks (NeurIPS 2017 · similarity 0.91)
- Learning to Generate Long-term Future via Hierarchical Prediction (ICML 2017 · similarity 0.88)
- Learning to Generate Time-Lapse Videos Using Multi-Stage Dynamic Generative Adversarial Networks (CVPR 2018 · similarity 0.88)
- Hierarchical Long-term Video Prediction without Supervision (ICML 2018 · similarity 0.86)
- Deep Learning in Latent Space for Video Prediction and Compression (CVPR 2021 · similarity 0.86)
- MCGAN: Enhancing GAN Training with Regression-Based Generator Loss (AAAI 2025 · similarity 0.83)
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.