enalis/LeafBenchV2

LeafBench 2.0

LeafBench 2.0 is a visual question answering (VQA) benchmark for evaluating fine-grained plant disease understanding in vision-language models (VLMs). Derived directly from LeafNet 2.0, the benchmark consists of multiple-choice questions spanning 9 complementary plant pathology tasks, designed to assess disease understanding beyond coarse category recognition.

LeafBench 2.0 was evaluated across 16 VLMs including 7 CLIP-based models, 7 generative VLMs, and 2 proprietary models (GPT-4o, Gemini 2.5 Pro), revealing substantial performance gaps between coarse recognition tasks and fine-grained pathological reasoning.

This benchmark accompanies the paper:

LeafNet 2.0: A Multiregional Image–Text Dataset for Vision-Language Modeling and Reasoning of Plant Diseases
Trang V. Nguyen, Khang Nguyen Quoc, David Harwath, Phuong D. Dao
The University of Texas at Austin · Korea University

---

Benchmark at a Glance

Property	Value
Benchmark images	6,000 (173 crop–disease classes)
Questions	52,187
Evaluation tasks	9
Answer format	Multiple choice (A / B / C / D)
Distractors per question	3
Source dataset	LeafNet 2.0 (216,815 images)
Models evaluated	16 (7 CLIP-based, 7 generative, 2 proprietary)
Best overall accuracy	67.78% (Gemini 2.5 Pro)
Best open-source accuracy	60.02% (Qwen3-VL-4B)

---

The 9 Benchmark Tasks

Task	Abbreviation	Description
Disease Identification	DI	Identify the disease present in the leaf image
Pathogen Classification	PC	Classify the causal pathogen type (fungus, bacterium, virus, etc.)
Crop Species Identification	CSI	Identify the crop species shown in the image
Symptom Identification	SI	Identify the specific visible symptom(s)
Healthy/Diseased Classification	HDC	Determine whether the leaf is healthy or diseased
Scientific Name Classification	SNC	Assign the correct scientific name to the disease or pathogen
Lesion Identification	LI	Identify lesion type, morphology, or distribution pattern
Leaf Symptom Detection	LSD	Detect the presence or location of symptomatic regions
Disease Severity Classification	DSC	Classify the severity level of the visible disease

Tasks range from coarse recognition (HDC, CSI) to fine-grained pathological reasoning (PC, SNC, LI), providing a comprehensive evaluation spectrum for plant disease AI systems.

---

Dataset Structure

leafbench2.0/
├── images/
│   ├── [crop1]-[disease1]/
│   │   ├── [crop1]-[disease1]-id_00001.png
│   │   ├── [crop1]-[disease1]-id_00002.png
│   │   └── ...
│   ├── [crop2]-[disease2]/
│   │   └── ...
│   └── ...
├── leafbenchv2.csv       ← full benchmark annotation file
└── LICENSE.txt

Images keep the [crop]-[disease]-id_[ID] name of the same image in LeafNet 2.0, so a benchmark image can be matched to its LeafNet 2.0 copy and caption.

leafbenchv2.csv columns

Column	Description
question_id	Question identifier ([task]:[n])
image_path	Relative path to the benchmark image
task	Task category (e.g., DI, PC, CSI, SI, HDC, SNC, LI, LSD, DSC)
question	Task-specific multiple-choice question
choice_a	Answer option A
choice_b	Answer option B
choice_c	Answer option C
choice_d	Answer option D
ground_truth	Correct answer label (A, B, C, or D)

---

Benchmark Results

Evaluation was conducted across 16 VLMs using classification accuracy per task. Results reveal clear specialization differences across model families.

The results below were computed on the original release (54,580 questions), which also contained 2,351 questions on Plant Pathology 2021 – FGVC8 and Kaggle wheat images. Those images cannot be redistributed and were removed in October 2026. In the same release, LeafNet 2.0 dropped byte-identical copies in Coffee-Healthy and Coffee-Miner: benchmark images that were copies of one another now share one file (100 → 63 images), and 42 questions that became exact repeats on the same image were removed.

Overall Accuracy (Top Models)

Model	Type	Overall Accuracy
Gemini 2.5 Pro	Proprietary	67.78%
GPT-4o	Proprietary	64.72%
Qwen3-VL-4B	Generative (open)	60.02%
AgriCLIP	CLIP-based (domain)	~65% (DI task)
SCOLD	CLIP-based (domain)	Competitive on SI tasks

Task-Level Highlights

Task	Easiest Model	Hardest for Most Models
Healthy/Diseased Classification (HDC)	GPT-4o: 93.50%	No — most models perform well
Crop Species Identification (CSI)	Gemini 2.5 Pro: 76.80%	Moderate difficulty
Leaf Symptom Detection (LSD)	Gemma4-8B: 92.76%	No
Pathogen Classification (PC)	All models struggle	Yes — requires subtle discrimination
Scientific Name Classification (SNC)	All models struggle	Yes — requires domain knowledge
Lesion Identification (LI)	All models struggle	Yes — subtle morphological cues

Key finding: Agriculture-adapted models (AgriCLIP, SCOLD) consistently outperformed several larger general-domain architectures on symptom-oriented tasks (SI, LSD), demonstrating the value of domain-specific pretraining. Most models achieved >90% on HDC but dropped substantially on PC, SNC, and LI, confirming that LeafBench 2.0 captures meaningful fine-grained complexity beyond coarse disease recognition.

---

Usage

Load benchmark annotations

from datasets import load_dataset

ds = load_dataset("enalis/LeafBenchV2")
print(ds["test"][0])
# → {
#     "question_id": "DI:1",
#     "image_path": "images/Apple-Black_rot/Apple-Black_rot-id_00156.png",
#     "task": "DI",
#     "question": "What disease is shown on the Apple leaf in this image?",
#     "choice_a": "Black rot",
#     "choice_b": "Powdery mildew",
#     "choice_c": "Narrow Brown Spot",
#     "choice_d": "leaf blight",
#     "ground_truth": "A"
#   }

Load with images for direct model evaluation

from huggingface_hub import snapshot_download
from PIL import Image

root = snapshot_download("enalis/LeafBenchV2", repo_type="dataset")
ds = ds.map(lambda x: {"image": Image.open(f"{root}/{x['image_path']}")})

Filter by task

pc_subset = ds["test"].filter(lambda x: x["task"] == "PC")

---

Evaluation Script

The official evaluation code, model implementations, and reproduction scripts are available at:

🔗 https://github.com/EnalisUs/LeafBench

Environment:

Python 3.10
torch==2.7.0
transformers==4.51.3
opencv-python==4.11.0.86
accelerate==1.8.1
torchvision==0.22.0
peft==0.15.0

---

Intended Use

LeafBench 2.0 is designed for:

• Fine-grained VLM evaluation — assessing plant disease understanding across 9 pathology tasks with varying difficulty levels.
• Agricultural domain adaptation benchmarking — comparing general-domain and agriculture-adapted models on symptom-level reasoning.
• Diagnostic reasoning research — studying whether multimodal models learn biologically meaningful symptom features or rely on superficial visual correlations.
• Zero-shot and few-shot evaluation — testing model generalization to unseen crop-disease combinations or geographic distributions.
• Multimodal reasoning studies — examining causal interpretation, uncertainty estimation, and differential diagnosis in plant pathology.

---

Relationship to LeafNet 2.0

LeafBench 2.0 is derived directly from the LeafNet 2.0 evaluation subset (6,000 images in this release), preserving the same variability in:

• Imaging devices and conditions (smartphones, digital cameras, controlled/natural backgrounds)
• Geographic and environmental diversity (9 regions)
• Disease severity and progression stages (early/late)
• Crop and disease coverage (173 crop–disease classes)

This ensures that benchmark performance reflects real-world agricultural conditions rather than idealized controlled settings.

---

Limitations

• The benchmark covers 9 tasks but does not include open-ended generation tasks (e.g., free-form caption generation or differential diagnosis). Future versions may extend to these settings.
• Performance on PC, SNC, and LI tasks is generally low across all current architectures, suggesting these remain open research challenges rather than solved problems.
• Task difficulty is inherently tied to disease and crop distribution in LeafNet 2.0; rare classes may be underrepresented in the benchmark.
• As with LeafNet 2.0, a small proportion of images exhibit ambiguous stage-specific features that may affect ground-truth reliability for the DSC task.

---

License

The benchmark annotations (questions, answer options, task labels and answers) are released under the Creative Commons Attribution 4.0 International (CC BY 4.0) license. The images are not relicensed: each keeps the license of its source dataset, as listed in LICENSE.txt (shared with LeafNet 2.0). Images from Plant Pathology 2021 – FGVC8 and the Kaggle wheat leaf dataset are not redistributed.