Latvian Sign Language Landmark Corpus
Identifikátory výsledku
Kód výsledku v IS VaVaI
<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216208%3A11320%2F26%3AD7P28V4G" target="_blank" >RIV/00216208:11320/26:D7P28V4G - isvavai.cz</a>
Výsledek na webu
<a href="https://repository.clarin.lv/repository/xmlui/handle/20.500.12574/139" target="_blank" >https://repository.clarin.lv/repository/xmlui/handle/20.500.12574/139</a>
DOI - Digital Object Identifier
—
Alternativní jazyky
Jazyk výsledku
angličtina
Název v původním jazyce
Latvian Sign Language Landmark Corpus
Popis výsledku v původním jazyce
The corpus contains MediaPipe-extracted landmark data representing 45 Latvian Sign Language signs. It includes 33 alphabet letters (a-z), 11 numbers (0-10), and a pause, which were captured from videos featuring several different signers. The collection covers both isolated signs and sign combinations forming complete words or short sentences. For each video frame, 543 landmarks were obtained using MediaPipe Holistic: pose (0-32), face (33-500), left hand (501-521), and right hand (522-542). The dataset is provided as .npy files containing the landmark data suitable for machine learning. A supplementary Jupyter notebook is also provided, showcasing example on how the landmarks can be loaded and visualized as a 2D human skeleton.
Název v anglickém jazyce
Latvian Sign Language Landmark Corpus
Popis výsledku anglicky
The corpus contains MediaPipe-extracted landmark data representing 45 Latvian Sign Language signs. It includes 33 alphabet letters (a-z), 11 numbers (0-10), and a pause, which were captured from videos featuring several different signers. The collection covers both isolated signs and sign combinations forming complete words or short sentences. For each video frame, 543 landmarks were obtained using MediaPipe Holistic: pose (0-32), face (33-500), left hand (501-521), and right hand (522-542). The dataset is provided as .npy files containing the landmark data suitable for machine learning. A supplementary Jupyter notebook is also provided, showcasing example on how the landmarks can be loaded and visualized as a 2D human skeleton.
Klasifikace
Druh
O - Ostatní výsledky
CEP obor
—
OECD FORD obor
10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)
Návaznosti výsledku
Projekt
—
Návaznosti
—
Ostatní
Rok uplatnění
2025
Kód důvěrnosti údajů
S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů