Research map / Video / Video segmentation and tracking
Recurrence over Video Frames (RoVF) for Animal Re-identification
Mitchell Rogers, Kobe Knowles, Gaël Gendron, Shahrokh Heidari, Isla Duporge, David Arturo Soriano Valdez, Mihailo Azhar, Padriac O’Leary, Simon Eyre, Michael Witbrock, Patrice Jean Delmas
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 Perceiver recurrent head over frozen DINOv2-Base (RoVF) | DINOv2-Base pretrained · Meerkat re-ID (Wellington Zoo clips) | Top-1: 38.5 -> 49.0 | Improved |
| add Perceiver recurrent head over frozen DINOv2-Base (RoVF) | DINOv2-Base pretrained · Meerkat re-ID (Wellington Zoo clips) | Top-3: 73.0 -> 64.5 | No clear effect |
| fine-tune DINOv2-Base on the meerkat triplet task (ablation) | DINOv2-Base pretrained · Meerkat re-ID (Wellington Zoo clips) | Top-1: 38.5 -> 29.0 | Hurt |
| fine-tune DINOv2-Small on the meerkat triplet task (ablation) | DINOv2-Small pretrained · Meerkat re-ID (Wellington Zoo clips) | Top-1: 41.5 -> 33.5 | Hurt |
| train RoVF for 10 epochs instead of 5 (ablation) | RoVF-Base · Meerkat re-ID (Wellington Zoo clips) | Top-1: 48.0 -> 49.0 | No clear effect |
| train fine-tuned DINOv2-Small 10 epochs instead of 5 (ablation) | DINOv2-Small fine-tuned · Meerkat re-ID (Wellington Zoo clips) | Top-1: 26.0 -> 33.5 | Improved |
| use video frames averaged from pretrained DINOv2-Small instead of single-frame ResNet-50 | ResNet-50 · Meerkat re-ID (Wellington Zoo clips) | Top-1: 26.5 -> 41.5 | Improved |
| scale pretrained DINOv2 from Small to Giant | DINOv2 · Meerkat re-ID (Wellington Zoo clips) | Top-1: 41.5 -> 42.0 | No clear effect |
Closest papers by meaning
- Active Learning for Animal Re-Identification with Ambiguity-Aware Sampling (AAAI 2026 · similarity 0.87)
- Region-Aware Multimodal Interleaving for Animal Re-Identification (ECCV 2026 · similarity 0.87)
- Species-Agnostic Patterned Animal Re-identification by Aggregating Deep Local Features (IJCV 2024 · similarity 0.86)
- Generating the Wild: Individual-Consistent Image-to-Video Generation for Wildlife (NeurIPS 2026 · similarity 0.86)
- Web-Scale Collection of Video Data for 4D Animal Reconstruction (NeurIPS 2025 · similarity 0.84)
- Animal-Bench: Benchmarking Multimodal Video Models for Animal-centric Video Understanding (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.
