Junwoo Cho, Seungtae Nam, Daniel Rho, Jong Hwan Ko, Eunbyung Park
TL;DR提出了可流式传输的神经场模型,通过可执行的各种宽度的子网络,可以重构不同品质和部分信号,例如,较小的子网络产生平缓和低频信号,而较大的子网络可以表示细节,实验结果表明,该方法有效地应用于 2D 图像,视频和 3D 信号。同时,该方法还利用参数共享来提高训练稳定性。
Abstract
neural fields have emerged as a new data representation paradigm and have
shown remarkable success in various signal representations. Since they preserve
signals in their network parameters, the data transfer by sending and receiving
the entire model parameters prevents this emerging t