List your company
ResearchFormerly WizWand

Articulatory-to-Acoustic Inversion (AAI) on EMA dataset unseen speakers (test)

0.7818CCACP-T
0.6641760.6947130.725250.755787Jul 2, 2026
Updated 4mo ago

Evaluation results

MethodLinks
ACP-TInputs=TERA, EMA data used for training (%)=752026.070.78181.36
ACP-TInputs=TERA, EMA data used for training (%)=1002026.070.7811.3619
BaselineInputs=TERA, EMA data used for training (%)=752026.070.77771.3774
ACP-TInputs=TERA, EMA data used for training (%)=502026.070.77551.3849
BaselineInputs=TERA, EMA data used for training (%)=1002026.070.77171.3878
ACP-TInputs=MFCC, EMA data used for training (%)=1002026.070.76891.3848
BaselineInputs=TERA, EMA data used for training (%)=502026.070.76641.386
ACP-TInputs=MFCC, EMA data used for training (%)=752026.070.76531.416
ACP-TInputs=TERA, EMA data used for training (%)=252026.070.76211.411
ACP-TInputs=MFCC, EMA data used for training (%)=502026.070.76161.416
BaselineInputs=MFCC, EMA data used for training (%)=1002026.070.75631.4143
ACP-TInputs=TERA, EMA data used for training (%)=12.52026.070.75621.411
ACP-TInputs=TERA, EMA data used for training (%)=6.252026.070.75611.3999
BaselineInputs=TERA, EMA data used for training (%)=252026.070.7541.416
BaselineInputs=MFCC, EMA data used for training (%)=502026.070.74881.4357
BaselineInputs=MFCC, EMA data used for training (%)=752026.070.74881.4357
ACP-TInputs=MFCC, EMA data used for training (%)=252026.070.74691.4514
ACP-TInputs=MFCC, EMA data used for training (%)=12.52026.070.73991.4779
BaselineInputs=TERA, EMA data used for training (%)=12.52026.070.73961.4515
BaselineInputs=TERA, EMA data used for training (%)=6.252026.070.73251.4788
BaselineInputs=MFCC, EMA data used for training (%)=252026.070.72651.4854
ACP-TInputs=MFCC, EMA data used for training (%)=6.252026.070.72591.5241
BaselineInputs=MFCC, EMA data used for training (%)=12.52026.070.69911.554
BaselineInputs=MFCC, EMA data used for training (%)=6.252026.070.66871.5992