Research map / Architectures and efficiency / Pruning and efficient tuning
LiDeRe: A Lightweight Readout for Fast and Data-Efficient Dense Prediction
Timo Lüddecke, Jan Frederik Meier, Jan van Delden, Alexander Ecker
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 |
|---|---|---|---|
| replace fixed bilinear with learnable interpolation prior (ablation) | LiDeRe (ViT-B) · PlantDoc detection | mAP: 41.2 -> 47.2 | Improved |
| replace fixed bilinear with learnable interpolation prior (ablation) | LiDeRe (ViT-B) · Horse10 pose | RMSE: 4.41 -> 3.26 | Improved |
| add learnable interpolation prior (ablation) | LiDeRe (ViT-B) · Camo | F-measure: 83.5 -> 82.7 | Mixed |
| add content-guided attention (ablation) | LiDeRe (ViT-B) · PlantDoc detection | mAP: 45.2 -> 47.2 | Improved |
| add all three modules (I, AC, FF) (ablation) | LiDeRe (ViT-B) · PlantDoc detection | mAP: 37.1 -> 47.2 | Improved |
| use LiDeRe readout instead of linear evaluation (ablation) | DINOv3 ViT-B · Horse10 pose | RMSE: 7.41 -> 3.26 | Improved |
| add upscaling transposed conv to linear evaluation (ablation) | linear readout · Horse10 pose | RMSE: 7.41 -> 8.92 | Hurt |
| add LoRA to the readout | LiDeRe (DINOv3 ViT-L) · Leaf Disease Segmentation | FG-IoU: 77.6 -> 81.1 | Improved |
| use DINOv3 instead of ImageNet pretraining | LiDeRe (ViT-B) · PlantDoc detection | mAP: 36.9 -> 47.2 | Improved |
| use DINOv3 ViT-L instead of DINOv2 ViT-L | LiDeRe · Horse10 pose | RMSE: 1.88 -> 2.42 | Hurt |
| use ViT-L instead of ViT-B | LiDeRe (DINOv3) · Pascal VOC | mIoU: 86.4 -> 88.9 | Improved |
| use LiDeRe instead of Conv-LoRA (training at 1024 px) | LiDeRe · training speed | iter/s: 1.22 -> 3.61 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- SAM-PARSER — SAM-PARSER: Fine-Tuning SAM Efficiently by Parameter Space Reconstruction (AAAI 2024) · 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
- VFM-Adapter: Adapting Visual Foundation Models for Dense Prediction with Dynamic Hybrid Operation Mapping (AAAI 2025 · similarity 0.86)
- Parameter Efficient Fine-tuning via Cross Block Orchestration for Segment Anything Model (CVPR 2024 · similarity 0.85)
- Lessons and Insights from a Unifying Study of Parameter-Efficient Fine-Tuning (PEFT) in Visual Recognition (CVPR 2025 · similarity 0.84)
- Point-PEFT: Parameter-Efficient Fine-Tuning for 3D Pre-trained Models (AAAI 2024 · similarity 0.84)
- Densely Connected Parameter-Efficient Tuning for Referring Image Segmentation (AAAI 2025 · similarity 0.84)
- LoKi: Low-dimensional KAN for Efficient Fine-tuning Image Models (CVPR 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.
