The FeatureTable after batch correction via percentile normalization is a FeatureTable[PercentileNormalized] table. It is not equivalent to the FeatureTable from denoising, as it contains percentile-normalized data (rather than read counts). The rows and columns are the same, but the values inside the table are represent percentiles of the distribution in the controls. Check out our publication to learn more about the method and the data type.
With respect to QIIME 2, this output feature table is not a compatible input to any QIIME 2 functions. In general, we envisioned this data to be used for non-parametric differential abundance testing. Any other analyses are at your own risk, and we encourage you to think about what this data means before just plugging in anything to any analyses.