Overview Downloads  Publications Performance Collaboration
Aurora Evaluations
Overview


On this web page we present a summary of the results for several training and testing conditions being evaluated on the ASL Fingerspelling dataset. Detailed explanations of the baseline system and the corresponding evaluation conditions are available on-line. In these evaluations, we will test the following conditions:
  • Singer dependent:Sample data of each signer is devided into 10 subset, 9 of them are used for training and the rest one is used for testing.

  • Half-Half: Half of the whole dataset is used for training and the other half is used for testing.

  • Leave-One-Out: Leave one subject for testing and use the rest subjects for training.



Performance Summary (Singer Dependent)
Set 1 2 3 4 5 6 7 8 9 10 Average
1 1.17% 1.92% 2.33% 1.17% 1.17% 2.00% 1.00% 2.00% 1.58% 2.25% 1.66%
1 1.58% 2.42% 2.25% 1.67% 1.17% 1.83% 1.58% 2.08% 1.42% 1.92% 1.79%
1 1.92% 2.17% 2.42% 2.25% 2.58% 2.08% 2.83% 2.25% 1.50% 3.08% 2.31%
1 4.25% 4.17% 4.42% 3.42% 3.25% 4.00% 4.08% 3.50% 4.50% 4.08% 3.97%
1 0.08% 0.33% 0.75% 0.75% 0.75% 0.92% 0.58% 0.17% 0.50% 0.50% 0.53%


Total average singer dependent error rate is 2.05%



Performance Summary (Half-Half)
Set 1 Set 2 Average
9.90% 5.78% 7.84%


Performance Summary (Leave-One-Out)
Set 1 Set 2 Set 3 Set 4 Set 5 Average
40.10% 39.86% 46.11% 66.33% 40.25% 46.53%

Footer

Up | Home | Courses | Projects | Proposals | Publications
Please direct questions or comments to joseph.picone@gmail.com