Implementation of histogram equalization, adaptive histogram equalization, histogram matching, image reconstruction from tie points, image resize, rotation, translation, shear, image negatives, log transformation, gamma transformation, contrast stretching, thresholding, intensity slicing and bitplane slicing using OpenCV C++ without using inbuilt functions.
Complie and Run
................
g++ -o a2_paras_kumar__2016csb1047 a2_paras_kumar_2016csb1047.cpp pkg-config opencv --cflags --libs
./a2_paras_kumar__2016csb1047
Comments:
1.Running the tie points subprogram requires input in format: x1 x4 y1 y4 x2 x5 y2 y5 x3 x6 y3 y6 (this line will be 0 0 0 0 in case of non translational correction)
where, INPUT POINTS OUTPUT POINTS (x1, y1) -> (x4 ,y4) (x2, y2) -> (x5 ,y5) (x3, y3) -> (x6 ,y6)
Note: *Example Data and image names are provided during run for easy testing.(I have included two images q.jpg and q1.jpg for this test)
2.Rmse is dispalyed for Resizing both bilinear and nearest neighbour, histogram equalization and histogram matching in rmse and psnr.
3.I have used same interpolation function for resize, shear and rotation. And the rmse in case of reisze comes around 8 for bilinear and 0 for nearest neighbour.
4.There are options for selecting type of interpolation in each geometric transformation.
5.Adaptive histogram takes long time for HD images, so try with poor quality images only.