数据集获取

git clone https://github.com/uwm-bigdata/Multi-modal-wound-classification-using-images-and-locations.git

curl -L -o repo.zip https://github.com/uwm-bigdata/Multi-modal-wound-classification-using-images-and-locations/archive/refs/heads/main.zip
unzip repo.zip

源站 README 摘录（使用方式）

Call for Papers: Special Issue on <ins>Recent Advances in Artificial Intelligence for Wound Assessment</ins>

In collaboration with Bioengineering, an open access journal by MDPI, we are soliciting papers in various AI technologies for wound assessment. More detail can be found below or at the journal’'s webpage.

Wound classification using images and locations

Dataset information: AZH dataset is collected over a two-year clinical period at the AZH Wound and Vascular Center in Milwaukee, Wisconsin. The dataset includes 730 wound images in .jpg format. The images are of various sizes, where the width ranging from 320 to 700 pixels and the height ranging from 240 to 525 pixels. These images contain four different wound types: venous, diabetic, pressure, and surgical. iPad Pro (software version 13.4.1) and a Canon SX 620 HS digital camera are used to capture the images, and labeling is done by a wound specialist from the AZH Wound and Vascular Center. For most images in our dataset, each image is taken from a separate patient. But there are a few cases where multiple photos were taken from the same patient at different body sites or various healing stages. For the latter case, the wound shapes are different, so they are considered separate images.
Pre-Processing:

1. The wound ROIs are cropped using our developed localizer (https://ieeexplore.ieee.org/abstract/document/9785640).
2. Wound professionals annotate the corresponding ROI locations with detailed location information using our custom body map (Download BodyMap.zip).
   Please read the following articles for more details.

Publication

1. Patel, Y., Shah, T., Mrinal Kanti Dhar, Zhang, T., Niezgoda, J., Gopalakrishnan, S., & Yu, Z. (2024). Integrated image and location analysis for wound classification: a deep learning approach. Scientific Reports, 14(1). https://doi.org/10.1038/s41598-024-56626-w
2. Anisuzzaman, D. M., Patel, Y., Rostami, B., Niezgoda, J., Gopalakrishnan, S., & Yu, Z. (2022). Multi-modal wound classification using wound image and location by deep neural network. Scientific Reports, 12(1). https://doi.org/10.1038/s41598-022-21813-0

数据加载示例（图像类）

from PIL import Image
import glob, os

files = (glob.glob(os.path.join(path, "**", "*.png"), recursive=True)
       + glob.glob(os.path.join(path, "**", "*.jpg"), recursive=True)
       + glob.glob(os.path.join(path, "**", "*.tif"), recursive=True))
print("图像文件数:", len(files))
img = Image.open(files[0]); print("尺寸/模式:", img.size, img.mode)

# torchvision Dataset 方式:
# from torchvision import datasets
# ds = datasets.ImageFolder(path)  # 要求 子目录=类别

目录组织与标注格式以源站说明和下载后实际文件为准。

完整仓库：github.com/uwm-bigdata/Multi-modal-wound-classification-using-images-and-locations