Affiliation:
1. Ocean University of China
2. The Hong Kong Polytechnic University
Abstract
We present an attention-weighted loss in a photometric stereo neural network to improve 3D surface recovery accuracy in complex-structured areas, such as edges and crinkles, where existing learning-based methods often failed. Instead of using a uniform penalty for all pixels, our method employs the attention-weighted loss learned in a self-supervise manner for each pixel, avoiding blurry reconstruction result in such difficult regions. The network first estimates a surface normal map and an adaptive attention map, and then the latter is used to calculate a pixel-wise attention-weighted loss that focuses on complex regions. In these regions, the attention-weighted loss applies higher weights of the detail-preserving gradient loss to produce clear surface reconstructions. Experiments on real datasets show that our approach significantly outperforms traditional photometric stereo algorithms and state-of-the-art learning-based methods.
Publisher
International Joint Conferences on Artificial Intelligence Organization
Cited by
16 articles.
订阅此论文施引文献
订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献
1. A Method of Micro-Geometric Details Preserving in Surface Reconstruction from Gradient;Proceedings of the 31st ACM International Conference on Multimedia;2023-10-26
2. A Critical Analysis of NeRF-Based 3D Reconstruction;Remote Sensing;2023-07-18
3. Learning Deep Photometric Stereo Network with Reflectance Priors;2023 IEEE International Conference on Multimedia and Expo (ICME);2023-07
4. Deep Discrete Wavelet Transform Network for Photometric Stereo;2023 24th International Conference on Digital Signal Processing (DSP);2023-06-11
5. Efficient Feature Fusion for Learning-Based Photometric Stereo;ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP);2023-06-04