In this paper, we propose a volume-to-volume fully convolutional network architecture effectively trained with hybrid loss, referred as HL-FCN, ...
確定! 回上一頁