208
The Media Pipeline library was utilize to extract features, which is beneficial in teaching
the model the necessary hand characteristics during signing. The use of pictorial models
captures key hand placement points, emphasizes small differences between characters, and
prioritizes the signer's hand over the background.
K points are establish in each frame, with K being the number of frames. Each identified
point is represented as
[ ]
[ ]
[ ]
, where i = 1, 2, 3, …, N, and j = 1, 2, 3, …, K. For
static gestures, all designated points should maintain their coordinates over time. This
applies to the dynamic six letters as well, as they are treat as static, and the last frames of
their respective videos are take [13]. The evaluation point,
[ ]
[ ]
[ ]
, bu yerda
[ ]
∑
[ ]
,
[ ]
∑
[ ]
, was employed. In our study, each hand displays 28
symbols.
Trening:
Transfer learning techniques offer a solution to the problems of learning and
computing data that are invisible to computers. This approach presents two main
advantages. Firstly, it avoids the challenge of training large data sets from the beginning,
and secondly, it reduces the high computational resource requirements.
