Comment by aub3bhat

Without any human assessment or blind testing, claiming one method outperforms another using an ex post facto unsupervised metric is not scientific.