Monsoon 2017: PGSSP Team 11
Steps for Executing SRCNN.py
- Download SRCNN.py, train.zip, test.zip, and val.zip in local folder.
- Extract train.zip, test.zip and val.zip in train, test and val folders respectively.
- Use python 2.7.14
- Install all dependent packages Keras, scipy, numpy, os, timeit, cv2.
- Open terminal/command prompt and navigate to above download folder.
- Execute command "python SRCNN.py"
- Execution will take maximum 30 mins.
- test folder will contain inuput low resolution images, bicubic images and predicted images.
- Best PSNR achieved will be reported at the end of execution.
Steps for Executing SRKNN.py
- Download SRKNN.py, train.zip and test.zip, in local folder.
- Extract train.zip and test.zip in train, test folders respectively.
- Use python 3.6.1
- Install all dependent packages sklearn, scipy, numpy, os, matplotlib.
- Open terminal/command prompt and navigate to above download folder.
- Execute command "python SRKNN.py "
- Execution will take less than 5 mins.
- Input high resolution, reconstructed image and blurred input images are saved in the output folder path given in the command line prompt.
- PSNR will be reported on the console after reconstruction of every image in the test folder given in the command line prompt.
Steps for Executing SRSVR.py
- Download SRSVR.py and SVR.zip in local folder.
- Extract SVR.zip to get train , test and Result folders respectively.
- Use python 2.7
- Install all dependent packages PIL, skimage,sklearn, scipy, numpy, glob, matplotlib.
- Open terminal/command prompt and navigate to above download folder.
- Execute command "python SRSVR.py . The Train , Test and Result folders should be in the same path. K value is 10e4
- Execution will take less than 5 mins.
- Result folder will have the Low resolution Image and the corresponding reconstructed Image with LR and SR prefixed.
- PSNR ( Individual and Mean) will be reported on the console after reconstruction of every image as well as summary in command prompt