Weijuan Hu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Weijuan Hu has a total of 42 co-authors. Here are the top co-authors by number of shared publications:
Jiafei Zhang
3 shared publications- Zhiguo Han
MetaPheno Laboratory, Shanghai, China
3 shared publications Binghui Xu
Wuhan Digital Engineering Institute, Wuhan, China
1 shared publicationBingpeng Yang
1 shared publication- Chang Chi
School of Artificial Intelligence, Wuhan University of Technology, Wuhan, Hubei, China
1 shared publication - Chang Liu
School of Artificial Intelligence, Wuhan University of Technology, Wuhan, Hubei, China
1 shared publication
Latest Publications
- Open Access
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 6661-6669
Segmentation with Explicit Uncertainty: Toward X-ray CT-Based Phenotyping of Living Soybean Seed Growth for Precision Breeding
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 6661-6669
Open AccessSegmentation with Explicit Uncertainty: Toward X-ray CT-Based Phenotyping of Living Soybean Seed Growth for Precision Breeding
- Open Access
- 1 August 2026
- Published by KEAI Publishing in New Crops
Artificial intelligence and digital phenotyping shape the future of plant breeding: evolution, challenges, and a roadmap from AI breeding 1.0 to 2.0
- 1 August 2026
- Published by KEAI Publishing in New Crops
Open AccessArtificial intelligence and digital phenotyping shape the future of plant breeding: evolution, challenges, and a roadmap from AI breeding 1.0 to 2.0
- Open Access
- 3 July 2026
- Published by Elsevier in Smart Agricultural Technology
- Vol. 14
NFStitch: Near-ground field image stitching of crops using gantry frames based on spatial layout priors and object protection
- 3 July 2026
- Published by Elsevier in Smart Agricultural Technology
- Vol. 14
Open AccessNFStitch: Near-ground field image stitching of crops using gantry frames based on spatial layout priors and object protection
- 15 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Selected Topics in Signal Processing
- Vol. 20 (2) , 129-141
DepthCropSeg++: Scaling a Crop Segmentation Foundation Model With Depth-Labeled Data
We introduce DepthCropSeg++, a foundation model for crop segmentation, capable of segmenting different crop species under open in-field environment. Crop segmentation is a fundamental task for modern agriculture, which closely relates to many downstream tasks such as plant phenotyping, density estimation, and weed control. In the era of foundation models, a number of generic large language and vision models have been developed. These models have demonstrated remarkable realworld generalization due to significant model capacity and largescale datasets. However, current crop segmentation models mostly learn from limited data due to expensive pixel-level labelling cost, often performing well only under specific crop types or controlled environment. In this work, we follow the vein of our previous work DepthCropSeg, an almost unsupervised approach to crop segmentation, to scale up a cross-species and crossscene crop segmentation dataset, with 28,406 images across 30+ species and 15 environmental conditions. We also build upon a state-of-the-art semantic segmentation architecture ViT-Adapter architecture, enhance it with dynamic upsampling for improved detail awareness, and train the model with a two-stage selftraining pipeline. To systematically validate model performance, we conduct comprehensive experiments to justify the effectiveness and generalization capabilities across multiple crop datasets. Results demonstrate that DepthCropSeg++ achieves $93.11\%$ mIoU on a comprehensive testing set, outperforming both supervised baselines and general-purpose vision foundation models like Segmentation Anything Model (SAM) by significant margins ($+0.36\%$ and +$48.57\%$ respectively). The model particularly excels in challenging scenarios including night-time environment ($86.90\%$ mIoU), high-density canopies ($90.09\%$ mIoU), and unseen crop varieties ($90.09\%$ mIoU), indicating a new state of the art for crop segmentation.
- 15 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Selected Topics in Signal Processing
- Vol. 20 (2) , 129-141
DepthCropSeg++: Scaling a Crop Segmentation Foundation Model With Depth-Labeled Data
We introduce DepthCropSeg++, a foundation model for crop segmentation, capable of segmenting different crop species under open in-field environment. Crop segmentation is a fundamental task for modern agriculture, which closely relates to many downstream tasks such as plant phenotyping, density estimation, and weed control. In the era of foundation models, a number of generic large language and vision models have been developed. These models have demonstrated remarkable realworld generalization due to significant model capacity and largescale datasets. However, current crop segmentation models mostly learn from limited data due to expensive pixel-level labelling cost, often performing well only under specific crop types or controlled environment. In this work, we follow the vein of our previous work DepthCropSeg, an almost unsupervised approach to crop segmentation, to scale up a cross-species and crossscene crop segmentation dataset, with 28,406 images across 30+ species and 15 environmental conditions. We also build upon a state-of-the-art semantic segmentation architecture ViT-Adapter architecture, enhance it with dynamic upsampling for improved detail awareness, and train the model with a two-stage selftraining pipeline. To systematically validate model performance, we conduct comprehensive experiments to justify the effectiveness and generalization capabilities across multiple crop datasets. Results demonstrate that DepthCropSeg++ achieves $93.11\%$ mIoU on a comprehensive testing set, outperforming both supervised baselines and general-purpose vision foundation models like Segmentation Anything Model (SAM) by significant margins ($+0.36\%$ and +$48.57\%$ respectively). The model particularly excels in challenging scenarios including night-time environment ($86.90\%$ mIoU), high-density canopies ($90.09\%$ mIoU), and unseen crop varieties ($90.09\%$ mIoU), indicating a new state of the art for crop segmentation.
- Open Access
- 4 December 2025
- Published by Elsevier in Smart Agricultural Technology
- Vol. 13
TraitDiscover: An automated high-throughput platform for multimodal plant phenotyping with real-time trait analysis
- 4 December 2025
- Published by Elsevier in Smart Agricultural Technology
- Vol. 13
Open AccessTraitDiscover: An automated high-throughput platform for multimodal plant phenotyping with real-time trait analysis
- Open Access
- 1 December 2025
- Published by KEAI Publishing in The Crop Journal
- Vol. 13 (6) , 1928-1938
TillerPET: High-throughput in-situ phenotyping of rice tiller number and compactness from post-harvest stubble
- 1 December 2025
- Published by KEAI Publishing in The Crop Journal
- Vol. 13 (6) , 1928-1938
Open AccessTillerPET: High-throughput in-situ phenotyping of rice tiller number and compactness from post-harvest stubble