AdaGaR：用於動態場景重建的自適應蓋博表徵法

摘要

基於單目影片重建動態三維場景需同時捕捉高頻外觀細節與時間連續運動。現有採用單一高斯基元的方法受其低通濾波特性限制，而標準Gabor函數會引入能量不穩定性。此外，時間連續性約束的缺失常導致插值過程出現運動偽影。我們提出AdaGaR——一個在顯式動態場景建模中同時解決頻率自適應性與時間連續性的統一框架。我們引入自適應Gabor表徵，通過可學習頻率權重與自適應能量補償擴展高斯模型，以平衡細節捕捉與穩定性。針對時間連續性，我們採用帶時間曲率正則化的三次埃爾米特樣條來確保平滑運動演化。結合深度估計、點追蹤與前景遮罩的自適應初始化機制，在訓練初期建立穩定的點雲分佈。在Tap-Vid DAVIS上的實驗顯示出最先進的性能（PSNR 35.49、SSIM 0.9433、LPIPS 0.0723），並在幀插值、深度一致性、影片編輯與立體視圖合成等任務中展現強泛化能力。項目頁面：https://jiewenchan.github.io/AdaGaR/

English

Reconstructing dynamic 3D scenes from monocular videos requires simultaneously capturing high-frequency appearance details and temporally continuous motion. Existing methods using single Gaussian primitives are limited by their low-pass filtering nature, while standard Gabor functions introduce energy instability. Moreover, lack of temporal continuity constraints often leads to motion artifacts during interpolation. We propose AdaGaR, a unified framework addressing both frequency adaptivity and temporal continuity in explicit dynamic scene modeling. We introduce Adaptive Gabor Representation, extending Gaussians through learnable frequency weights and adaptive energy compensation to balance detail capture and stability. For temporal continuity, we employ Cubic Hermite Splines with Temporal Curvature Regularization to ensure smooth motion evolution. An Adaptive Initialization mechanism combining depth estimation, point tracking, and foreground masks establishes stable point cloud distributions in early training. Experiments on Tap-Vid DAVIS demonstrate state-of-the-art performance (PSNR 35.49, SSIM 0.9433, LPIPS 0.0723) and strong generalization across frame interpolation, depth consistency, video editing, and stereo view synthesis. Project page: https://jiewenchan.github.io/AdaGaR/

AdaGaR：用於動態場景重建的自適應蓋博表徵法

AdaGaR: Adaptive Gabor Representation for Dynamic Scene Reconstruction

摘要

Support