siddharthdhara17 commited on
Commit
3912f4b
·
verified ·
1 Parent(s): 64e32b5

Updated README.md and dataset card

Browse files
Files changed (2) hide show
  1. README.md +13 -0
  2. dataset_card.json +10 -0
README.md ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # 🩺 LIDC-IDRI Text-Annotated
2
+
3
+ We release a **text-annotated version of the LIDC-IDRI dataset**, where each annotation is **carefully curated from structured metadata** provided by radiologists (e.g., malignancy, size, shape, margin, texture, spiculation, etc.).
4
+
5
+ This enables new research directions in:
6
+ - **Multi-modal learning** (image + text)
7
+ - **Text-guided medical image segmentation**
8
+
9
+ ## Includes
10
+ - Radiologist Nodule annotations (radiologist contours, malignancy scores)
11
+ - Natural language descriptions derived from metadata
12
+
13
+ > This dataset is based on the publicly available LIDC-IDRI release from TCIA. Text annotations are derived from radiologist-provided metadata.
dataset_card.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "pretty_name": "LIDC-IDRI Text-Annotated",
3
+ "language": ["en"],
4
+ "license": "cc-by-nc-sa-4.0",
5
+ "task_categories": ["image-segmentation", "image-to-text"],
6
+ "source_datasets": ["LIDC-IDRI (TCIA)"],
7
+ "multilinguality": "monolingual",
8
+ "size_categories": ["10K<n<100K"],
9
+ "dataset_summary": "A text-augmented version of the LIDC-IDRI dataset with natural language annotations curated from structured radiologist metadata. Useful for multi-modal learning in medical imaging."
10
+ }