Table 1.
Overview of the mentioned approaches.
Table 2.
Signs used for classification, extracted from LSA64 dataset.
Fig 1.
The pipeline followed in the presented approach.
Fig 2.
Hand landmarks obtained with MediaPipe.
Fig 3.
Example of hand landmarks obtained for a sign sequence.
Fig 4.
Preprocessing of the set of signals of a video.
Fig 5.
Boxplot of variances of different projection vectors, by class.
Table 3.
Used classifiers and their parameters.
Table 4.
Configuration of the classification.
Table 5.
Obtained results with different configurations.
Table 6.
Obtained results for each parameter value.
Table 7.
Mean accuracy values obtained with the best configuration (RGB and B/W color spaces) for each class.
Table 8.
Statistics of results obtained with best parameter settings.