Research map / Video / Video segmentation and tracking
Latent-INR: A Flexible Framework for Implicit Representations of Videos with Discriminative Semantics
Shishira R Maiya, Anubhav Gupta, Matthew A Gwilliam, Max Ehrlich, Abhinav Shrivastava
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 |
|---|---|---|---|
| add CLIP alignment loss on latents (lambda 1e-3 to 1) (ablation) | Latent-INR · MSR-VTT | T2V R@1: 0.1 -> 30.2 | Improved |
| increase CLIP loss weight lambda from 0 to 1.0 (ablation) | Latent-INR · MSR-VTT | Reconstruction PSNR: 30.03 -> 28.61 | Hurt |
| increase CLIP lambda beyond 0.01 (ablation) | Latent-INR · MSR-VTT | T2V R@1: 30.2 -> 29.7 / 30.2 | No clear effect |
| use latent-space interpolation through hypernetwork instead of time or weight interpolation of other INRs | video INR · Bunny alpha=8 | PSNR: 18.67 -> 25.88 (vs NIRVANA) | Improved |
| retrieve with Latent-INR latents aligned to CLIP instead of CLIP features | retrieval · COIN segment-level | R@1: 6.60 -> 6.40 | No clear effect |
| retrieve with Latent-INR latents instead of CLIP features | retrieval · HowTo100m subset segment-level | R@10: 40.85 -> 48.60 | Improved |
| use Fourier features instead of hash-grid positional encoding (ablation) | Latent-INR · UVG compression | training speed / reconstruction: up to 30% faster training | Mixed |
| use larger patch sizes in patch-centroid model (ablation) | Latent-INR · UVG | PSNR | Hurt |
| increase latent dimension beyond 512 (ablation) | Latent-INR · UVG bosphore | PSNR vs bitrate | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- NIRVANA — NIRVANA: Neural Implicit Representations of Videos With Adaptive Networks and Autoregressive Patch-Wise Modeling (CVPR 2023) · better on 2 of 2 shared results
- NeRV — NeRV: Neural Representations for Videos (NeurIPS 2021) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Boosting Neural Representations for Videos with a Conditional Decoder (CVPR 2024 · similarity 0.88)
- Combining Frame and GOP Embeddings for Neural Video Representation (CVPR 2024 · similarity 0.88)
- NVRC: Neural Video Representation Compression (NeurIPS 2024 · similarity 0.87)
- One Token per Highly Selective Frame: Towards Extreme Compression for Long Video Understanding (NeurIPS 2025 · similarity 0.87)
- Bias for Action: Video Implicit Neural Representations with Bias Modulation (CVPR 2025 · similarity 0.87)
- Explaining the Implicit Neural Canvas: Connecting Pixels to Neurons by Tracing their Contributions (CVPR 2024 · similarity 0.87)
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.
