Research map / Time series / Foundation models for time series
PhenoLeaf-TS: A Time-Series Benchmark for Leaf Instance Segmentation, Tracking, and Growth Stage Classification
Rijad Saric, Basim Azam, Sarmad Khan, Edhem Custovic
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 |
|---|---|---|---|
| fine-tune from PhenoLeaf-TS weights instead of zero-shot | Mask R-CNN (IS-3) · Komatsuna | mAP: 34.7 -> 86.1 (+51.4) | Improved |
| fine-tune from PhenoLeaf-TS weights instead of zero-shot | Mask R-CNN (IS-3) · Ara2012 | mAP: 55.3 -> 73.9 (+18.6) | Improved |
| fine-tune from PhenoLeaf-TS weights instead of zero-shot | Mask R-CNN (IS-3) · Ara2013 | mAP: 53.5 -> 75.2 (+21.7) | Improved |
| fine-tune from PhenoLeaf-TS weights instead of zero-shot | Mask2Former (IS-8) · Komatsuna | mAP: 47.1 -> 76.7 (+29.6) | Improved |
| fine-tune from PhenoLeaf-TS weights instead of zero-shot | Swin-T (CL-6) · Ara2013 classification | accuracy: 21.0 -> 78.0 (+57.0) | Improved |
| fine-tune from PhenoLeaf-TS weights instead of zero-shot | Swin-T (CL-6) · Ara2012 classification | accuracy: 100.0 -> 100.0 | No clear effect |
| add text prompting via Grounding DINO (SAM3) over unprompted SAM2 | SAM2 (IS-7) · PhenoLeaf-TS segmentation | mAP: 55.0 -> 71.1 | Improved |
| increase Mask R-CNN backbone from R50 to R101 (ablation) | Mask R-CNN · PhenoLeaf-TS segmentation | mAP: 73.2 -> 72.4 | Hurt |
| increase Mask R-CNN backbone from R50 to X101 (ablation) | Mask R-CNN · PhenoLeaf-TS segmentation | mAP: 73.2 -> 71.3 | Hurt |
| add cascade refinement to Mask R-CNN R50 (ablation) | Mask R-CNN R50-FPN · PhenoLeaf-TS segmentation | mAP: 73.2 -> 71.9 | Hurt |
| use ResNet-101 instead of ResNet-50 for growth stage classification (ablation) | ResNet · PhenoLeaf-TS classification | accuracy: 89.4 -> 89.7 | No clear effect |
| use Swin-T instead of ViT-B-16 for growth stage classification | ViT-B-16 · PhenoLeaf-TS classification | accuracy: 85.1 -> 91.7 | Improved |
Closest papers by meaning
- Joint Multi-Leaf Segmentation, Alignment, and Tracking for Fluorescence Plant Videos (TPAMI 2017 · similarity 0.90)
- PlantTraitNet: An Uncertainty-Aware Multimodal Framework for Global-Scale Plant Trait Inference from Citizen Science Data (AAAI 2026 · similarity 0.88)
- Plant Taxonomy Meets Plant Counting: A Fine-Grained, Taxonomic Dataset for Counting Hundreds of Plant Species (CVPR 2026 · similarity 0.87)
- GrowFields: Compositional 4D Neural Fields for Topology-Changing Plant Growth (ECCV 2026 · similarity 0.86)
- BioCLIP: A Vision Foundation Model for the Tree of Life (CVPR 2024 · similarity 0.83)
- Utilizing Vision-Language Models for Detection of Leaf-Based Diseases in Tomatoes (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.
