We compare the performances of pre-trained OCR and segmentation-free approach for speech text of comic books written in Latin script.
確定! 回上一頁