Downloads · 30 days
12
4% of all-time downloads
pneumonia-cxr
Dimaodessa/pneumonia-cxr
pneumonia-cxr is a image classification model from Dimaodessa. Use it when you need a label for an image. It is set up for transformers.
This model performs binary classification and segmentation for pneumonia (lung opacity) in frontal chest radiographs. It is a tfefficientnetv2s backbone with a U-Net decoder and linear classification head. The model w…
All-time downloads
327
Public
Parameters
22.2M
89.1 MB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors89.1 MB · 100%
At a glance
- Task
- Image Classification
- Library
- transformers
- Model type
- pneumonia
- Access
- Public
- Created
- Mar 12, 2026
- Updated
- Mar 12, 2026
- SHA
- 026cee21
Base models
Model card
From the Hugging Face model README
This model performs binary classification and segmentation for pneumonia (lung opacity) in frontal chest radiographs.
It is a tf_efficientnetv2_s backbone with a U-Net decoder and linear classification head.
The model was trained on the RSNA Pneumonia Detection Challenge dataset and the SIIM-FISABIO-RSNA COVID-19 Detection dataset.
Both of these datasets were annotated with bounding boxes, which were converted to ellipsoid segmentation masks.
Classification performance on a holdout test set of 1,334 images from the RSNA dataset and 317 images from the SIIM-FISABIO-RSNA dataset:
RSNA + SIIM-FISABIO-RSNA (n=1,651): AUC 0.900
RSNA (n=1,334): AUC 0.885
SIIM-FISABIO-RSNA (n=317) : AUC 0.914
- Task
- Image Classification
- Library
- transformers
- Type
- pneumonia
- Created
- Mar 12, 2026
- Updated
- Mar 12, 2026