Tomato Multi-Angle Multi-Pose Dataset for Fine-Grained Phenotyping

2026-02-28 · Source: Machine learning : nature.com subject feeds · Field: Agriculture & Food Systems — Artificial Intelligence & Machine Learning, Precision Agriculture & Smart Farming, Crop Science & Plant Technology · Depth: Advanced, medium

Summary

The Tomato Multi-Angle Multi-Pose (TomatoMAP) dataset is introduced to enhance the accuracy and reproducibility of fine-grained plant phenotyping for Solanum lycopersicum. This comprehensive dataset comprises 68,080 RGB images, including 3,616 high-resolution macrophotographs (3648 × 5472) with semantic annotations and 64,464 moderate-resolution images (1080 × 1440) captured from 12 plant poses at four camera elevations. Each image features manually annotated bounding boxes for seven regions of interest (leaves, panicle, flower clusters, fruit clusters, axillary shoot, shoot, and whole-plant area) and labels for 50 BBCH growth stages. The dataset supports real-time applications, with models like MobileNetv3, YOLOv11, and Mask R-CNN benchmarked for accuracy, mAP, and inference FPS. AI models trained on TomatoMAP achieved accuracy comparable to five domain experts, with reliability supported by Cohen’s Kappa statistics.

Key takeaway

For Computer Vision Engineers developing agricultural AI, TomatoMAP provides a robust, publicly available dataset to train and validate models for fine-grained tomato phenotyping. You should explore its multi-angle imagery and detailed annotations to improve the accuracy and reproducibility of your plant analysis systems, especially for real-time applications. The dataset's benchmarking against models like MobileNetv3 and YOLOv11 offers a strong baseline for your own model selection and optimization.

Key insights

TomatoMAP dataset enables accurate, reproducible fine-grained tomato phenotyping using multi-angle imagery and AI models.

Principles

• Multi-angle, multi-pose imaging reduces observer bias.
• Semantic annotation improves model accuracy.
• AI models can match human expert phenotyping.

Method

The dataset construction involves capturing images from 12 plant poses at four camera elevations, followed by manual annotation of seven regions of interest and 50 BBCH growth stages, then benchmarking AI models.

In practice

• Utilize TomatoMAP for tomato growth stage classification.
• Apply bounding box annotations for organ detection.
• Benchmark real-time models like YOLOv11 for efficiency.

Topics

• Plant Phenotyping
• TomatoMAP Dataset
• Object Detection Models
• Agricultural AI
• Semantic Annotation

Code references

• 0YJ/TomatoMAP
• mamta-joshi-gehlot/Tomato-Village
• up2metric/tomatOD
• spMohanty/PlantVillage-Dataset
• laboroai/LaboroTomato

Best for: Computer Vision Engineer, Research Scientist, AI Researcher, AI Scientist, Data Scientist

Related on AIssential

• FruitEnsemble: MLLM-Guided Arbitration for Heterogeneous ensemble in Fine-Grained Fruit Recognition — A two-stage dynamic inference framework combines heterogeneous ensembles with MLLM-guided arbitration for fine-grained fruit recognition.
• 🪝Syn4D: Multiview Synthetic 4D Dataset🪝 👉Syn4D is novel multi-view synthetic dataset of... — Syn4D is a synthetic multi-view 4D dataset offering comprehensive ground-truth for dynamic scene analysis.
• USU-Corn-WeedDB: A UAV RGB Image Dataset for Multi-Species Weed Detection in Forage Corn — A new UAV RGB dataset enables multi-species weed detection in forage corn, supporting supervised and semi-supervised deep learning.
• LatticeVision: Image to Image Networks for Modeling Non-Stationary Spatial Data — Image-to-image networks globally estimate non-stationary spatial parameters faster and more accurately than local methods.
• Meta AI Releases Sapiens2: A High-Resolution Human-Centric Vision Model for Pose, Segmentation, Normals, Pointmap, and Albedo — Sapiens2 unifies multiple human vision tasks into a single model, improving performance and simplifying pipelines.
• Unlocking Multi-Spectral Data for Multi-Modal Models with Guided Inputs and Chain-of-Thought Reasoning — Generalist LMMs can interpret multi-spectral data zero-shot by converting it to pseudo-images and using detailed, CoT-enhanced prompts.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Machine learning : nature.com subject feeds.