🎉 Congratulations on Paper Acceptance at ACCV 2024!

This work introduces a simple yet effective encoder-decoder framework that encodes positional and viewpoint coordinates into a Shared Feature Field (SFF), which is then decoded into an implicit Signed Distance Field (SDF) and a color field. By employing a weight-sharing encoder, the joint optimization of SDF and color fields is significantly enhanced, allowing for better utilization of limited scene information. Furthermore, the integration of a periodic sine function as an activation function eliminates the need for a positional encoding layer and greatly reduces surface rippling artifacts. Empirical results demonstrate that this method surpasses existing techniques in reconstructing textureless and minimally visible surfaces, producing higher-quality novel views, and achieving superior multi-view reconstruction with fewer input images.

All accepted papers are available at: https://openaccess.thecvf.com/ACCV2024.

Please visit the introduction page for more detailed information about this paper.