Skip to content
Open
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion notebooks/high_dimensional_ct_hugging_face.ipynb
Original file line number Diff line number Diff line change
Expand Up @@ -280,7 +280,7 @@
"\n",
"The MedGemma image encoder accepts RGB, 8-bits per channel images as input. MedGemma has also been trained to interpret CT images where the RGB channels of input images correspond to a novel windowing representation to enable it to interpret multiple representations of CT imaging simultaneously. Specifically, MedGemma 1.5 has been trained with the components defined as follows:\n",
" * Red (component 0): Wide window; range: -1024 HU ([air](https://en.wikipedia.org/wiki/Hounsfield_scale)) to 1024 HU ([above bone](https://en.wikipedia.org/wiki/Hounsfield_scale))\n",
" * Green (component 1): Soft tissue window; range: 135 HU ([fat](https://en.wikipedia.org/wiki/Hounsfield_scale)) to 215 HU ([start of bone](https://en.wikipedia.org/wiki/Hounsfield_scale))\n",
" * Green (component 1): Soft tissue window; range: -135 HU ([fat](https://en.wikipedia.org/wiki/Hounsfield_scale)) to 215 HU ([start of bone](https://en.wikipedia.org/wiki/Hounsfield_scale))\n",
" * Blue (component 2): Brain window; range: 0 HU ([water](https://en.wikipedia.org/wiki/Hounsfield_scale)) to 80 HU ([brain](https://en.wikipedia.org/wiki/Hounsfield_scale))\n",
"\n",
" Since each of the RGB channels in the prompt images corresponds to a windowing prompt, imaging prepared using this method will visually appear in color.\n",
Expand Down