Table 1 Baseline characteristics for the multimodal training dataset comprising chronic wound images with corresponding patient metadata and weakly supervised ground truth masks. Note that in two cases it was not possible to exactly identify the wound type - “arterial or venous” and “venous or pressure”. \(^*\) - with fungal component.
Category | Total | Category | Total |
|---|---|---|---|
No. of wound images | 1142 | No. of venous or pressure wound images | 1 |
No. of DFU wound images | 1111 | No. of patients | 308 |
No. of venous wound images | 13 | No. of appointments | 94 |
No. of arterial wound images | 12 | No. of male patients | 229 |
No. of pressure wound images | 1 | No. of female patients | 79 |
No. of dermatoliposclerosis wound images | 1 | Median patient age | 70 |
No. of bacterial infection wound images\(^*\) | 1 | Median male patient age | 69 |
No. of ulcer on necrobiosis lipoidica wound images | 1 | Median female patient age | 70 |
No. of arterial or venous wound images | 1 |
