Xianlin Song, Jiaqing Dong, Minghao Liu, Zehao Sun, Zibang Zhang, Jianghao Xiong, Zilong Li, Xuan Liu, Qiegen Liu*
Real-time intelligent 3D holographic photography for real-world scenarios
Optics Express Vol. 32, Issue 14, pp. 24540-24552 (2024)
https://doi.org/10.1364/OE.529107
Visualization.3.mp4
This code runs with Python 3.8.17, Pytorch 2.0.1 and TensorRT 8.6.0
- ./src/
- train.py: The training code of the model
- NET1.py: The network structure of the model1
- dataLoader.py: The data loader of the model
- rtholo.py: The code of the real-time holography
- predict_rgbd_multiprocess.py: The testing code of the model
- trt.py: The code of TensorRT class
- getBlaze.py: This code is for generating a blazed grating
- GCD_ctrl.py: This code is for controlling the motorized linear stage
- depthcamera_ctrl.py: This code is for controlling the depth camera Realsense D435
- gxipy: The SDK of the Daheng camera
- ./trt/
- trt_create_v1.py: This code is used to generate the TRT model
- trt_inference_v1.py: This code is used to test the TRT model
python ./src/train.py --p_loss --l2_loss --num_epochs 60 --data_path <The address of your training set>
python predict_rgbd_multiprocess.py
We provide pretrained checkpoints. The pre-trained models in - ./src/checkpoints/CNN_1024_30/53.pth
We are thankful for the open source of tensor_holography ,HoloEncoder, HoloEncoder-Pytorch-Version and Self-Holo. These works are very helpful for our research.
-
Lens-less imaging via score-based generative model
[Paper] [Code] -
Multi-phase FZA Lensless Imaging via Diffusion Model
[Paper] [Code] -
Imaging through scattering media via generative diffusion model
[Paper] [Code] -
High-resolution iterative reconstruction at extremely low sampling rate for Fourier single-pixel imaging via diffusion model
[Paper] [Code] -
Dual-domain Mean-reverting Diffusion Model-enhanced Temporal Compressive Coherent Diffraction Imaging
[Paper] [Code]