Datasets:

xuxuxuxuxu
/

SurgWound

Tasks:

Question Answering

Text Generation

Modalities:

Formats:

Languages:

Size:

Tags:

Libraries:

License:

Dataset card Data Studio Files Files and versions Community

xuxuxuxuxu commited on 7 days ago

Commit

d36a817

·

verified ·

1 Parent(s): 6de60be

Update README.md

Files changed (1) hide show

README.md +9 -12

README.md CHANGED Viewed

@@ -98,23 +98,20 @@ Using these images and their corresponding annotations, we design two tasks:
 The following is an example from SurgWound-Bench VQA Task:
 ```
-"id": "606.jpg_wound_location",
-"image_name": "606.jpg",
-"field": "Wound Location",
 "task_type": "multi_choice",
 "image": base64encoded_image,
-"question": "What is the wound location of this surgical wound?",
 "options": [
-    "Abdomen",
-    "Patella",
-    "Ankle",
-    "Facial region",
-    "Manus",
-    "Cervical region",
-    "Other",
     "Uncertain"
 ],
-"answer": "Abdomen"
 ```
 The following is an example from SurgWound-Bench Report Generation Task:

 The following is an example from SurgWound-Bench VQA Task:
 ```
+"id": "76.jpg_closure_method",
+"image_name": "76.jpg",
+"field": "Closure Method",
 "task_type": "multi_choice",
 "image": base64encoded_image,
+"question": "What is the closure method of this surgical wound?",
 "options": [
+    "Invisible",
+    "Sutures",
+    "Staples",
+    "Adhesives",
     "Uncertain"
 ],
+"answer": "Sutures"
 ```
 The following is an example from SurgWound-Bench Report Generation Task: