# -*- coding: iso-8859-1 -*- """ MoinMoin - Daltonize ImageCorrection - Effect Daltonize image correction algorithm implemented according to http://scien.stanford.edu/class/psych221/projects/05/ofidaner/colorblindness_project.htm Many thanks to Onur Fidaner, Poliang Lin and Nevran Ozguven for their work on this topic and for releasing their complete research results to the public (unlike the guys from http://www.vischeck.com/). This is of great help for a lot of people! Please note: Daltonize ImageCorrection needs * Python Image Library (PIL) from http://www.pythonware.com/products/pil/ * NumPy from http://numpy.scipy.org/ You can call Daltonize from the command-line with "daltonize.py C:\image.png" Explanations: * Normally this module is called from Moin.AttachFile.get_file * @param filename, fpath is the filename/fullpath to an image in the attachment dir of a page. * @param color_deficit can either be - 'd' for Deuteranope image correction - 'p' for Protanope image correction - 't' for Tritanope image correct Idea: * Since daltonizing an image takes quite some time and we don't want visually impaired users to wait so long until the page is loaded, this module has a command-line option built-in which could be called as a separate process after a file upload of a non visually impaired user in "AttachFile", e.g "spawnlp(os.NO_WAIT...)" * "AttachFile": If an image attachment is deleted or overwritten by a new version please make sure to delete the daltonized images and redaltonize them. * But all in all: Concrete implementation of ImageCorrection needs further thinking and discussion. This is only a first prototype as proof of concept. @copyright: 2007 by Oliver Siemoneit @license: GNU GPL, see COPYING for details. """ import os.path def execute(filename, fpath, color_deficit='d'): modified_filename = "%s-%s-%s" % ('daltonize', color_deficit, filename) head, tail = os.path.split(fpath) # Save transformed image to the cache dir #head = head.replace('attachments', 'cache') modified_fpath = os.path.join(head, modified_filename) # Look if requested image is already available if os.path.isfile(modified_fpath): return (modified_filename, modified_fpath) helpers_available = True try: import numpy from PIL import Image except: helpers_available = False if not helpers_available: return (filename, fpath) # Get image data im = Image.open(fpath) if im.mode in ['1', 'L']: # Don't process black/white or grayscale images return (filename, fpath) im = im.copy() im = im.convert('RGB') RGB = numpy.asarray(im, dtype=float) # Transformation matrix for Deuteranope (a form of red/green color deficit) lms2lmsd = numpy.array([[1,0,0],[0.494207,0,1.24827],[0,0,1]]) # Transformation matrix for Protanope (another form of red/green color deficit) lms2lmsp = numpy.array([[0,2.02344,-2.52581],[0,1,0],[0,0,1]]) # Transformation matrix for Tritanope (a blue/yellow deficit - very rare) lms2lmst = numpy.array([[1,0,0],[0,1,0],[-0.395913,0.801109,0]]) # Colorspace transformation matrices rgb2lms = numpy.array([[17.8824,43.5161,4.11935],[3.45565,27.1554,3.86714],[0.0299566,0.184309,1.46709]]) lms2rgb = numpy.linalg.inv(rgb2lms) # Daltonize image correction matrix err2mod = numpy.array([[0,0,0],[0.7,1,0],[0.7,0,1]]) # Get the requested image correction if color_deficit == 'd': lms2lms_deficit = lms2lmsd elif color_deficit == 'p': lms2lms_deficit = lms2lmsp elif color_deficit == 't': lms2lms_deficit = lms2lmst else: return (filename, fpath) # Transform to LMS space LMS = numpy.zeros_like(RGB) for i in range(RGB.shape[0]): for j in range(RGB.shape[1]): rgb = RGB[i,j,:3] LMS[i,j,:3] = numpy.dot(rgb2lms, rgb) # Calculate image as seen by the color blind _LMS = numpy.zeros_like(RGB) for i in range(RGB.shape[0]): for j in range(RGB.shape[1]): lms = LMS[i,j,:3] _LMS[i,j,:3] = numpy.dot(lms2lms_deficit, lms) _RGB = numpy.zeros_like(RGB) for i in range(RGB.shape[0]): for j in range(RGB.shape[1]): _lms = _LMS[i,j,:3] _RGB[i,j,:3] = numpy.dot(lms2rgb, _lms) ## # Save simulation how image is perceived by a color blind ## for i in range(RGB.shape[0]): ## for j in range(RGB.shape[1]): ## _RGB[i,j,0] = max(0, _RGB[i,j,0]) ## _RGB[i,j,0] = min(255, _RGB[i,j,0]) ## _RGB[i,j,1] = max(0, _RGB[i,j,1]) ## _RGB[i,j,1] = min(255, _RGB[i,j,1]) ## _RGB[i,j,2] = max(0, _RGB[i,j,2]) ## _RGB[i,j,2] = min(255, _RGB[i,j,2]) ## simulation = _RGB.astype('uint8') ## im_simulation = Image.fromarray(simulation, mode='RGB') ## simulation_filename = "%s-%s-%s" % ('daltonize-simulation', color_deficit, filename) ## simulation_fpath = os.path.join(head, simulation_filename) ## im_simulation.save(simulation_fpath) # Calculate error between images error = (RGB-_RGB) # Daltonize ERR = numpy.zeros_like(RGB) for i in range(RGB.shape[0]): for j in range(RGB.shape[1]): err = error[i,j,:3] ERR[i,j,:3] = numpy.dot(err2mod, err) dtpn = ERR + RGB for i in range(RGB.shape[0]): for j in range(RGB.shape[1]): dtpn[i,j,0] = max(0, dtpn[i,j,0]) dtpn[i,j,0] = min(255, dtpn[i,j,0]) dtpn[i,j,1] = max(0, dtpn[i,j,1]) dtpn[i,j,1] = min(255, dtpn[i,j,1]) dtpn[i,j,2] = max(0, dtpn[i,j,2]) dtpn[i,j,2] = min(255, dtpn[i,j,2]) result = dtpn.astype('uint8') # Save daltonized image im_converted = Image.fromarray(result, mode='RGB') im_converted.save(modified_fpath) return (modified_filename, modified_fpath) if __name__ == '__main__': import sys print "Daltonize image correction for color blind users" if len(sys.argv) != 2: print "Calling syntax: daltonize.py [fullpath to image file]" print "Example: daltonize.py C:/wikiinstance/data/pages/PageName/attachments/pic.png" sys.exit(1) if not (os.path.isfile(sys.argv[1])): print "Given file does not exist" sys.exit(1) extpos = sys.argv[1].rfind(".") if not (extpos > 0 and sys.argv[1][extpos:].lower() in ['.gif', '.jpg', '.jpeg', '.png', '.bmp', '.ico', ]): print "Given file is not an image" sys.exit(1) path, fname = os.path.split(sys.argv[1]) print "Please wait. Daltonizing in progress..." colorblindness = { 'd': 'Deuteranope', 'p': 'Protanope', 't': 'Tritanope',} for col_deficit in ['d', 'p', 't']: print "Creating %s corrected version" % colorblindness[col_deficit] modified_filename, modified_fpath = execute(fname, sys.argv[1], col_deficit) if modified_fpath == sys.argv[1]: print "Error while processing image: PIL/NumPy missing and/or source file is a grayscale image." else: print "Image successfully daltonized"