avana/sagoma
2
1
Fork 1
Code Issues Pull requests Releases Wiki Activity
Detect book shapes and more. Very machine learning.
39 commits 2 branches 0 tags 35 KiB
C++ 76.7%
Cuda 12.3%
CMake 5.4%
Python 5.3%
Shell 0.3%
Find a file
boyska fc6cffe2ba python cv3 compatibility
2017-02-23 03:13:15 +01:00
.gitignore fix release build 2017-01-15 02:05:27 +01:00
bs_dsp.cu powered by spray paint 2017-01-05 13:40:59 +01:00
CMakeLists.txt optimize further_point_from_line 2017-01-20 16:33:54 +01:00
cvutils.cpp FIX find_longest_line on edge case 2017-02-23 01:27:59 +01:00
cvutils.h memory fixes 2017-01-12 16:25:29 +01:00
edge_finder.py python cv3 compatibility 2017-02-23 03:13:15 +01:00
geometry.cpp Inizio trasposizione linea verticale 2017-02-23 00:19:04 +01:00
geometry.h Inizio trasposizione linea verticale 2017-02-23 00:19:04 +01:00
getfiles powered by spray paint 2017-01-05 13:40:59 +01:00
lines.cpp improve transformation: vertical points is better than corners 2017-02-23 01:28:47 +01:00
lines.h similar_fit much improved 2017-01-06 04:12:34 +01:00
README.md readme 2017-01-05 14:00:14 +01:00
TestChangeCoord.h FIXED and tested coordinate system change 2017-01-12 22:03:11 +01:00
TestLine.h geometry: creates line (in implicit form) 2017-01-12 18:20:55 +01:00

README.md

Misc tools to identify important points/lines of book photos.

Build howto

cmake && make
./getfiles

(getfiles will sync the files/ directory from a remote server; no big files on git!)

Please note that bs_dsp requires CUDA libraries to be built (and an NVidia card to be run). If you don't have the appropriate libraries, it will just be skipped.

Tools

bs_dsp

correct colors, contrast, etc

lines

Given an image like files/masckera.png it will extract the shape of the book