Update README.md

README.md

@@ -46,7 +46,7 @@ dataset_summary: '
 
   # Note: other available arguments include ''max_samples'', etc
 
-  dataset = load_from_hub("harpreetsahota/PlatSeg-Test")
+  dataset = load_from_hub("Voxel51/PlatSeg-Test")
 
 
   # Launch the App
@@ -60,10 +60,7 @@ dataset_summary: '
 
 # Dataset Card for PlantSeg_Test
 
-<!-- Provide a quick summary of the dataset. -->
-
-
-
+![image/png](plantseg.gif)
 
 
 This is a [FiftyOne](https://github.com/voxel51/fiftyone) dataset with 1200 samples.
@@ -84,141 +81,118 @@ from fiftyone.utils.huggingface import load_from_hub
 
 # Load the dataset
 # Note: other available arguments include 'max_samples', etc
-dataset = load_from_hub("harpreetsahota/PlatSeg-Test")
+dataset = load_from_hub("Voxel51/PlatSeg-Test")
 
 # Launch the App
 session = fo.launch_app(dataset)
 ```
 
+# Dataset Card for PlantSeg
 
 ## Dataset Details
 
 ### Dataset Description
 
-<!-- Provide a longer summary of what this dataset is. -->
-
-
+PlantSeg is a large-scale in-the-wild dataset for plant disease segmentation, containing 11,458 images with high-quality segmentation masks across 115 disease categories and 34 plant types. Unlike existing plant disease datasets that are collected in controlled laboratory settings, PlantSeg primarily comprises real-world field images with complex backgrounds, various viewpoints, and different lighting conditions. The dataset also includes an additional 8,000 healthy plant images categorized by plant type.
 
-- **Curated by:** [More Information Needed]
-- **Funded by [optional]:** [More Information Needed]
-- **Shared by [optional]:** [More Information Needed]
+- **Curated by:** Tianqi Wei, Zhi Chen, Xin Yu, Scott Chapman, Paul Melloy, and Zi Huang
+- **Shared by:** The University of Queensland; CSIRO Agriculture and Food
 - **Language(s) (NLP):** en
-- **License:** [More Information Needed]
+- **License:** CC BY-NC-ND 4.0
 
 ### Dataset Sources [optional]
 
-<!-- Provide the basic links for the dataset. -->
-
-- **Repository:** [More Information Needed]
-- **Paper [optional]:** [More Information Needed]
-- **Demo [optional]:** [More Information Needed]
+- **Repository:** https://doi.org/10.5281/zenodo.13293891
+- **Paper [optional]:** arXiv:2409.04038
 
 ## Uses
 
-<!-- Address questions around how the dataset is intended to be used. -->
-
 ### Direct Use
 
-<!-- This section describes suitable use cases for the dataset. -->
-
-[More Information Needed]
-
-### Out-of-Scope Use
-
-<!-- This section addresses misuse, malicious use, and uses that the dataset will not work well for. -->
+- Training and benchmarking semantic segmentation models for plant disease detection
+- Developing automated disease diagnosis systems for precision agriculture
+- Image classification for plant disease identification
+- Evaluating segmentation algorithms on in-the-wild agricultural imagery
+- Supporting integrated disease management (IDM) decision-making tools
+## Dataset Structure
 
-[More Information Needed]
+The dataset is organized as follows:
 
-## Dataset Structure
+- **images/**: Plant disease images in JPEG format
+- **annotations/**: Segmentation labels in PNG format (grayscale, where diseased pixels have class index values and background is zero)
+- **json/**: Original LabelMe annotation files in JSON format
+- **PlantSeg-Meta.csv**: Metadata file containing image name, plant type, disease type, resolution, label file path, mask ratio, source URL, and train/test split assignment
 
-<!-- This section provides a description of the dataset fields, and additional information about the dataset structure such as criteria used to create the splits, relationships between data points, etc. -->
+**Statistics:**
+- Total images: 11,458 diseased plant images + 8,000 healthy plant images
+- Disease categories: 115
+- Plant types: 34
+- Train/test split: 80/20 (stratified by disease type)
 
-[More Information Needed]
+**Plant categories are organized into four socioeconomic groups:**
+- Profit crops (e.g., Coffee, Tobacco): 9 diseases across 3 plants
+- Staple crops (e.g., wheat, corn, potatoes)
+- Fruits (e.g., apples, oranges): 39 diseases across 10 plants
+- Vegetables (e.g., tomatoes): 45 diseases across 15 plants
 
 ## Dataset Creation
 
 ### Curation Rationale
 
-<!-- Motivation for the creation of this dataset. -->
+Existing plant disease datasets are insufficient for developing robust segmentation models due to three key limitations:
 
-[More Information Needed]
+1. **Annotation Type:** Most datasets only contain class labels or bounding boxes, lacking pixel-level segmentation masks
+2. **Image Source:** Many datasets contain images from controlled laboratory settings with uniform backgrounds, which do not reflect real-world field conditions
+3. **Scale:** Existing segmentation datasets are small and cover limited host-pathogen relationships
 
-### Source Data
+PlantSeg addresses these gaps by providing the largest in-the-wild plant disease segmentation dataset with expert-validated annotations.
 
-<!-- This section describes the source data (e.g. news text and headlines, social media posts, translated sentences, ...). -->
+### Source Data
 
 #### Data Collection and Processing
 
-<!-- This section describes the data collection and processing process such as data selection criteria, filtering and normalization methods, tools and libraries used, etc. -->
+Images were collected using plant disease names as keywords from multiple internet sources:
+- Google Images
+- Bing Images
+- Baidu Images
 
-[More Information Needed]
+This multi-source collection strategy ensured geographic diversity, with images sourced from websites worldwide. After collection, a rigorous data cleaning process was conducted where annotators reviewed each image and removed incorrect or ambiguous images, with cross-validation by at least two annotators and expert review for discrepancies.
 
 #### Who are the source data producers?
 
-<!-- This section describes the people or systems who originally created the data. It should also include self-reported demographic or identity information for the source data creators if this information is available. -->
-
-[More Information Needed]
+Images were sourced from websites globally, representing diverse geographic regions, environmental conditions, and imaging setups. The original photographers/sources are not individually identified, but source URLs are preserved in the metadata for reproducibility and copyright compliance.
 
 ### Annotations [optional]
 
-<!-- If the dataset contains annotations which are not part of the initial data collection, use this section to describe them. -->
-
 #### Annotation process
 
-<!-- This section describes the annotation process such as annotation tools used in the process, the amount of data annotated, annotation guidelines provided to the annotators, interannotator statistics, annotation validation, etc. -->
-
-[More Information Needed]
+1. **Standard establishment:** A segmentation annotation standard was created to ensure consistent labeling of disease-affected areas
+2. **Annotator training:** Annotators were trained on the standard and required to annotate 10 test images for evaluation before proceeding
+3. **Annotation tool:** LabelMe (V5.5.0) was used for polygon annotation
+4. **Annotation guidelines:**
+   - Distinct lesions: annotated with individual polygons
+   - Overlapping lesions: annotated as combined affected areas
+   - Small clustered symptoms (rust, powdery mildew): meticulously annotated to reflect disease distribution
+   - Disease-induced deformities: also annotated
+5. **Quality control:** Each image subset was annotated by one annotator, then reviewed by another annotator, with final review by expert plant pathologists
 
 #### Who are the annotators?
 
-<!-- This section describes the people or systems who created the annotations. -->
-
-[More Information Needed]
-
-#### Personal and Sensitive Information
-
-<!-- State whether the dataset contains data that might be considered personal, sensitive, or private (e.g., data that reveals addresses, uniquely identifiable names or aliases, racial or ethnic origins, sexual orientations, religious beliefs, political opinions, financial or health data, etc.). If efforts were made to anonymize the data, describe the anonymization process. -->
+- 10 trained annotators who passed qualification evaluations
+- Supervised by two expert plant pathologists who established standards, evaluated annotator work, and performed final reviews
 
-[More Information Needed]
 
-## Bias, Risks, and Limitations
-
-<!-- This section is meant to convey both technical and sociotechnical limitations. -->
-
-[More Information Needed]
-
-### Recommendations
-
-<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
-
-Users should be made aware of the risks, biases and limitations of the dataset. More information needed for further recommendations.
-
-## Citation [optional]
-
-<!-- If there is a paper or blog post introducing the dataset, the APA and Bibtex information for that should go in this section. -->
+## Citation
 
 **BibTeX:**
-
-[More Information Needed]
+```bibtex
+@article{wei2024plantseg,
+  title={PlantSeg: A Large-Scale In-the-wild Dataset for Plant Disease Segmentation},
+  author={Wei, Tianqi and Chen, Zhi and Yu, Xin and Chapman, Scott and Melloy, Paul and Huang, Zi},
+  journal={arXiv preprint arXiv:2409.04038},
+  year={2024}
+}
+```
 
 **APA:**
-
-[More Information Needed]
-
-## Glossary [optional]
-
-<!-- If relevant, include terms and calculations in this section that can help readers understand the dataset or dataset card. -->
-
-[More Information Needed]
-
-## More Information [optional]
-
-[More Information Needed]
-
-## Dataset Card Authors [optional]
-
-[More Information Needed]
-
-## Dataset Card Contact
-
-[More Information Needed]
+Wei, T., Chen, Z., Yu, X., Chapman, S., Melloy, P., & Huang, Z. (2024). PlantSeg: A Large-Scale In-the-wild Dataset for Plant Disease Segmentation. arXiv preprint arXiv:2409.04038.