Ask Your Question

Revision history [back]

click to hide/show revision 1
initial version

Detect stamps (seals) on image

Dear forum,

I have a lot of images of documents, where I need to detect stamps. I couldn't make a solution that would automatically adjust to different kinds of images.

So here is my offer. I am offering $100 to anyone who can write a piece of python code (and share it with everyone here), that would scan the following images for stamps and return only the cropped part of the image with the stamp itself. The stamp must be visible fully.

http://data26.i.gallery.ru/albums/smsdocs/1-67c5e-36856--hash89960bf2.jpg (soft) http://data27.i.gallery.ru/albums/smsdocs/1-28661-66087--hash6246fd59.jpg (photo) http://data26.i.gallery.ru/albums/smsdocs/1-19450-34976--hash13b374d9.jpg (photo) http://data27.i.gallery.ru/albums/smsdocs/1-b3c5d-58851--hashec50bbdd.jpg (positioned incorrectly) http://data27.i.gallery.ru/albums/smsdocs/1-3eafb-67617--hash4f934d43.jpg (flares) http://data27.i.gallery.ru/albums/smsdocs/1-2a5aa-61913--hash72cf5ee3.jpg (scan marks)

Detect stamps (seals) on image

Dear forum,

I have a lot of images of documents, where I need to detect stamps. I couldn't make a solution that would automatically adjust to different kinds of images.

So here is my offer. I am offering $100 to anyone who can write a piece of python code (and share it with everyone here), that would scan the following images for stamps and return only the cropped part of the image with the stamp itself. The stamp must be visible fully.

http://data26.i.gallery.ru/albums/smsdocs/1-67c5e-36856--hash89960bf2.jpg (soft) http://data27.i.gallery.ru/albums/smsdocs/1-28661-66087--hash6246fd59.jpg (photo) http://data26.i.gallery.ru/albums/smsdocs/1-19450-34976--hash13b374d9.jpg (photo) http://data27.i.gallery.ru/albums/smsdocs/1-b3c5d-58851--hashec50bbdd.jpg (positioned incorrectly) http://data27.i.gallery.ru/albums/smsdocs/1-3eafb-67617--hash4f934d43.jpg (flares) http://data27.i.gallery.ru/albums/smsdocs/1-2a5aa-61913--hash72cf5ee3.jpg (scan marks)