Neural Field Classifiers via Target Encoding and Classification Loss
This work addresses a fundamental but overlooked question in neural fields for computer vision and graphics, offering a novel approach that could enhance robustness in tasks like novel view synthesis and geometry reconstruction.
The authors tackled the problem of whether classification models are better than regression models for neural field methods, and they proposed a Neural Field Classifier (NFC) framework that transforms existing neural field regressors into classification tasks, showing impressive effectiveness with nearly free extra computational costs.
Neural field methods have seen great progress in various long-standing tasks in computer vision and computer graphics, including novel view synthesis and geometry reconstruction. As existing neural field methods try to predict some coordinate-based continuous target values, such as RGB for Neural Radiance Field (NeRF), all of these methods are regression models and are optimized by some regression loss. However, are regression models really better than classification models for neural field methods? In this work, we try to visit this very fundamental but overlooked question for neural fields from a machine learning perspective. We successfully propose a novel Neural Field Classifier (NFC) framework which formulates existing neural field methods as classification tasks rather than regression tasks. The proposed NFC can easily transform arbitrary Neural Field Regressor (NFR) into its classification variant via employing a novel Target Encoding module and optimizing a classification loss. By encoding a continuous regression target into a high-dimensional discrete encoding, we naturally formulate a multi-label classification task. Extensive experiments demonstrate the impressive effectiveness of NFC at the nearly free extra computational costs. Moreover, NFC also shows robustness to sparse inputs, corrupted images, and dynamic scenes.