Table 2.
Recall and IOU of the ingredient prediction task for 10,000 samples.
Models | K = 1 | K = 5 | K = 10 | |||
---|---|---|---|---|---|---|
CVG (0.47) | IOU (0.28) | CVG (0.56) | IOU (0.38) | CVG (0.59) | IOU (0.40) | |
JE | 0.0679 | 0.0367 | 0.0622 | 0.0311 | 0.0591 | 0.0308 |
JE + SR | 0.0763 | 0.0371 | 0.0721 | 0.0365 | 0.0695 | 0.0362 |
Ki-Cook | 0.0743 | 0.0399 | 0.0736 | 0.0368 | 0.0700 | 0.0359 |
Ki-Cook + SSWC | 0.0783 | 0.0406 | 0.0742 | 0.0377 | 0.0709 | 0.0365 |
Ki-Cook + SSWOC | 0.0793 | 0.0414 | 0.0747 | 0.0386 | 0.0743 | 0.0383 |
As the ground truth is removed, the upper bound (maximum possible accuracy) for each run is as mentioned within the parentheses.
The bold values represent the highest CVG and IOU for a given column.