Research map / Vision-language and multimodal / CLIP and zero-shot

Utilizing Vision-Language Models for Detection of Leaf-Based Diseases in Tomatoes

AAAI 2025 · Vision-language and multimodal › CLIP and zero-shot · approach: Test-time adaptation of VLMs · 0 citations (OpenAlex)

Read the paperCheck your idea against 108,967 papers

AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Proposes (without experiments) using a fine-tuned vision-language model, combining a multi-stream CNN visual encoder with text embeddings of disease labels, to detect tomato leaf diseases.

Problem

Early detection of tomato leaf diseases (early blight, late blight, septoria leaf spot) to protect yields, since manual diagnosis is slow and error-prone.

Approach

Proposes a multi-modal pipeline that fuses multi-stream CNN image features with text embeddings of disease labels in a joint space and fine-tunes a pre-trained VLM (e.g., CLIP, VisualBERT) for image-text matching-based disease classification.

Key results

• No experimental results are reported; the paper describes evaluation criteria (accuracy, precision, recall, F1) but gives no numbers.
• Cites prior work Trivedi et al. (2021) reporting 98.49% accuracy for a CNN on a 3000-image tomato dataset, not the proposed method.

Limitations

• No experiments or quantitative results are reported; the approach is a proposal.
• Dataset is described only as 'large' and no specific dataset is named.
• VLMs are computationally demanding and performance depends on training-data quality; misclassification for similar symptoms.

Experimental setup

Metrics: accuracy precision recall F1-score

Builds on: CLIP

Closest papers by meaning

• AgMMU: A Comprehensive Agricultural Multimodal Understanding Benchmark (NeurIPS 2025 · similarity 0.84)
• Cropper: Vision-Language Model for Image Cropping through In-Context Learning (CVPR 2025 · similarity 0.81)
• Empowering and Assessing the Utility of Large Language Models in Crop Science (NeurIPS 2024 · similarity 0.81)
• Combining Deep Learning and Street View Imagery to Map Smallholder Crop Types (AAAI 2024 · similarity 0.81)
• AgroOmni: A Large-Scale Multi-view Agricultural Dataset for Cross-Scale Multimodal Reasoning (NeurIPS 2026 · similarity 0.80)
• VLM4Bio: A Benchmark Dataset to Evaluate Pretrained Vision-Language Models for Trait Discovery from Biological Images (NeurIPS 2024 · similarity 0.80)