SRILF Examples

All Examples

·        Examples on FRGC v1 with 2-fold cross-validation

·        Examples on FRGC v2 with 2-fold cross-validation

·        Examples on FRGC v2 using a model trained on FRGC v1

 

SRILF Examples on FRGCv2

(2-fold cross validation)

In this experiment we target 14 landmarks on 4007 facial scans from FRGCv2, split into training and test set by 2-fold cross validation, obtaining an average accuracy of 3.55 mm (error over all landmarks with respect to manual ground truth). We provide below a set of examples, separated into 3 categories:

·         The 10 cases with lowest overall errors

·         Average performing cases (the 20 cases with overall errors closest to the average accuracy obtained in this experiment)

·         The 10 cases with highest overall errors

 

The 10 cases with lowest errors (between 1.60 and 1.88 mm)

93

679

960

1201

1742

2009

2658

2788

2963

3067

 

Average performing cases (overall errors ~ 3.55 mm)

140

591

653

1037

1242

1266

1460

1490

1674

1696

1713

1925

2198

2234

2321

3007

3174

3308

3746

3870

 

The 10 cases with highest errors (between 7.73 and 11.90 mm)

146

381

517

862

982

1395

2642

2712

2899

3348

 

 

All Examples

·        Examples on FRGC v1 with 2-fold cross-validation

·        Examples on FRGC v2 with 2-fold cross-validation

·        Examples on FRGC v2 using a model trained on FRGC v1