Face Morphing Print-Scan Research

This is the process that is followed to develop high quality printed and scanned photos using scripts found in the Code folder

Presentations

https://pages.nist.gov/ifpc/2022/presentations/29_NIST-IFPC_CITeR_Morph_Presentation_Nov_17_2022.pdf

Equipment

Scanner
Epson Perfection V850 Pro

Printer
Canon Pixma Pro 100

Software

Adobe Photoshop
Visual Studio Code
PyCharm IDE (optional)
Scanner Driver and EPSON Scan Utility
Canon Pixma Pro 100 Driver

Requirements

See Requirements.txt for all python libraries used
For JavaScript Code run in debugging and have ExtendScript and ExtendScript Debugger on Visual Studio Code

Running Scripts and Processing

Preprocessing Stage - Set up pictures for printing

1. Resize digital morphs to 600 by 600 pixels using cropResize.py
   Edit Line 4 to the set of images that need to be printed

fileRoot = r"C:\Users\xxxx\imgs_to_resize//"

2. Arrange images on page for printing - Create Print Pages.py Edit Line 9 to the set of images that need to be printed
   Edit Line 10 to the location you want the Print Pages to be stored
   Edit Line 11 to fit the dataset that you are printing and image type: Bonafide or Morph

imgRoot = r"C://Users//xxxx//Dataset//imgs//"
printPageRoot = r"C://Users//xxxx//Dataset_Print_Pages//"
pageName = "Dataset_Morph_Print_Page_" or pageName = "Dataset_Bonafide_Print_Page_"

Edit Line 54 to fit the dataset. This is for the physical page numbering system

# Default Line
draw.text((2200, 3200), pageCounterStringTwo + " Data PS", (0, 0, 0), font=font)
# Use for Morphed Images
draw.text((2200, 3200), pageCounterStringTwo + " Data MPS", (0, 0, 0), font=font)
# Use for Bonafide Images
draw.text((2200, 3200), pageCounterStringTwo + " Data BPS", (0, 0, 0), font=font)

Prep for Printing Stage

3. Launch Adobe Photoshop
   3i. Navigate to print under File then Print with any file opened
   3ii. Set settings to image seen here
   You will have to modify the ICC to match the specific type of paper that is being used to print
   For help see here Default for printing should be:

Canon Pro-100 <GL><PP> 1/2 Photo Paper Plus Glossy&Gold  

Unless other types of paper are being used then adjust the ICC profile accordingly to match.

Begin Printing

4. Launch Visual Studio Code and open print.js
   4i. Run print.js with Run --> Start Debugging)
   4ii. {Select Host Application --> Adobe Photoshop]()
   4iii. Navigate to Adobe Photoshop and select folder containing pages for printing and press select folder

Scanning Stage

Prepare Scanner

5. Launch EPSON Scan
   5i. Load professional mode
   5ii Set adjustments to off (unchecked)
   5iii. Set ICM in configuration tab to sRGB
   Go to file configuration and assure:
   Image type: jpg or jpeg
   Image compression: 1, highest quality
   Folder location: respective dataset ex. Dataset_Bonafide_Page_PS_ or Dataset_Morph_Page_PS_
   Scan numbering: set to the respective page number that is being scanned

Processing Stage

5. Launch python IDE and open batch_process.py
   Edit lines 5 through lines 13

line 5 - line 8 uncomment
arg = 1
scanRoot = r"C:\Users\xxxxx\images_to_process//"
saveRoot1 = r"C:\Users\xxxxx\images_splilt//"
saveRoot2 = r"C:\Users\xxxxx\images_to_crop//"

and comment 
#arg = 2
#scanRoot = r"C:\Users\xxxxx\images_cropped//"
#saveRoot1 = r"C:\Users\xxxxx\images_cleaned//"
#saveRoot2 = saveRoot1

Line 5 and 10 determine which set of instructions are performed
scanRoot is the location of files to be processed
saveRoot1 is the save location of the files after the first pass
saveRoot1is the save location of the files after the second pass

6. Launch Visual Studio Code and open crop.js
   6i. Run crop.js with Run --> Start Debugging)
   6ii. {Select Host Application --> Adobe Photoshop]()
   6iii. Navigate to Adobe Photoshop and select folder containing pages for cropping and press select folder

Check for mis-cropped images in cropped folder - check with eyes and use sort by dimensions in file explorer
Isolate mis-cropped images and hand crop images as requireed

7. Launch python IDE and open batch_process.py
   Edit lines 5 through lines 13

line 5 - line 8 comment
#arg = 1
#scanRoot = r"C:\Users\xxxxx\images_to_process//"
#saveRoot1 = r"C:\Users\xxxxx\images_splilt//"
#saveRoot2 = r"C:\Users\xxxxx\images_to_crop//"

and uncomment 
arg = 2
scanRoot = r"C:\Users\xxxxx\images_cropped//"
saveRoot1 = r"C:\Users\xxxxx\images_cleaned//"
saveRoot2 = saveRoot1

scanRoot is the location cropped images from crop.js
saveRoot is the output of the cleaned images

8. Launch python IDE and open Batch Rename.py
   Edit lines 3 and 4

fileRoot = r"C:\Users\xxxx\files_to_rename//"
nameRoot = r"C:\Users\xxxx\original_files//"

nameRoot is the original files and are ufor naming are from the original source images
Copy and rename the cleaned folder from the previous step to dataset_Renamed

9. Launch python IDE and open cropResize.py
   Edit line 4

fileRoot = r"C:\Users\xxxx\imgs_to_resize//"

Copy and rename the renamed folder from the previous step to dataset_Resized