DoNotSnap

An experiment in detecting DoNotSnap badges in photos, to protect privacy.

This program allows you to detect and identify DoNotSnap badges via a sliding-window decision tree classifier (custom heuristics are used to reduce search space). The classifier is trained by matching samples against image templates using Affine-transform invariant SURF features.

You can find examples of using the classifier in classify.py and training a new classifier in train.py

A pre-trained classifier can be found in classifier.pkl Alternative versions of the same classifier are in classifier_alt_1.pkl and classifier_alt_2.pkl

Running classification

Run python classify.py <path-image-to-be-tested> This will deserialize the classifier from classifier.pkl and run it on the image you supplied. A sample image could be found in sample.jpg

Training your own classifier

Run python train.py <output-file> <total-number-of-samples> This will read the sample filenames from positive.txt and negative.txt files. Templates filenames are specified in templates.txt. A sample template could be found in template.png The output is a <output-file>.pkl with serialized classifier.

Dependencies

• opencv
• numpy
• sklearn
• matplotlib
• PIL

DoNotSnap sticker templates

In the repository you can find templates for printing DoNoSnap stickers.

avery_8254.pdf contains design compatible with US format paper. Compatible Avery templates:

• 15664
• 18664
• 45464
• 48264
• 48464
• 48864
• 5164
• 5264
• 55164
• 5524
• 55264
• 55364
• 55464
• 5664
• 58164
• 58264
• 8164
• 8254
• 8464
• 8564
• 15264
• 95940
• 95905

A-0004-01_P.pdf contains design compatible with A4-format paper. Compatible Avery templates:

• J8165
• J8165-10
• J8165-25
• J8165-40
• J8165-100
• J8165-250
• J8165-500
• J8365
• J8565
• J8565-25
• L7165
• L7165-40
• L7165-100
• L7165-250
• L7165-500
• LR7165-100
• L7165X
• L7165X-100
• L7165X-250
• L7565
• L7565-25
• L7965
• L7965-100
• L7993
• L7993-25
• LR7165
• LR7165-100