2020-01-09
We made some new fine adjustments, detail in link https://github.com/Deepyanyuan/DITMnet-ldr2hdr and paper "DITMnet: Learning to Reconstruct HDR Image Based on a Single-shot Filtered LDR Image". This paper has just been submiteted to IEEE journal translation on multimedia.
2019-7-1
HybridNet：ldr2hdr via CNNs, detail in paper 《HybridNet: Deep Inverse Tone Mapping Learning to Reconstruct HDR Image》
original HDR dataset come from online: DML-HDR {http://dml.ece.ubc.ca/data/DML-HDR/}, Fairchild-HDR {http://rit-mcsl.org/fairchild//HDRPS/HDRthumbs.html}, and Funt-HDR {https://www2.cs.sfu.ca/~colour/data/}.
pipeline:
step 0: generate training pairs by generate_train_pairs.py
step 1: create network, in this paper, we create a multi-branch and multi-ouput CNNs network, detail in network.py
step 2: train this network by HybridNet_train.py. Note that in this paper, we use three datasets, the first two had been training and the last dataset (Funt-HDR) had no training.
step 3: test this network by HybridNet_test.py.
step 4: predict the results if input any LDR image by HybridNet_predict.py
step 5: performance comparison by Matlab code
2019-9-01
supplement
Added a complete datasets and pre-trained parameters, linked as follows
linked：https://pan.baidu.com/s/18Ho7er1eF8YMKNDfPPiRFQ Extraction code：lbbd
Note that after downloading, you need to extract all the compressed files first, the default path can be