Table 1.
A breakdown of the subsets used to develop ALFA.
| Dataset | 1 | 2 | 3 | 4 | |||
|---|---|---|---|---|---|---|---|
| Experiment | Segmentation Network | Validity (Control-line status) | IgG Status Classification 1 | IgG Status Classification 2 | |||
| Purpose | Develop | Test | Test | Develop | Test | Develop | Test |
| No. of samples | 415 | 83 | 187 | 864 | 294 | 1699 | 237 |
| No. of IgG positives | – | – | – | 351 | 143 | 641 | 79 |
| No. of invalids | – | – | 12 | – | – | – | – |
| The subsets from REACT-2 Study-5 Rounds which were combined to form the datasets (“Included” indicates inclusion) | |||||||
| Round-1, Sample A | Included | – | Included | Included | |||
| Round-1, Sample B | Included | Included | Included | Included | |||
| Round-1, Sample C | – | – | Included | Included | |||
| Round-2 sample | – | – | Included | Included | |||
| Round-5 sample | – | – | – | Included | |||
“Experiment” refers to which component of ALFA was being developed/tested. Most datasets were split into development and test subsets and indicated in the “purpose” row. The origins of the images are also shown in the bottom half of the table. “Round” refers to the round of the REACT-2 Study-5 from which the data were collected. “Included” indicate whether the sample was included in the dataset.