src/methods/Chardon_tom/reconstruct.py

@@ -7,7 +7,7 @@ Students can call their functions (declared in others files of src/methods/your_
 import numpy as np
 
 from src.forward_model import CFA
-from src.methods.chardon_tom.utils import *
+from src.methods.Chardon_tom.utils import *
 import pywt
 
 #!!!!!!!! It is normal that the reconstructions lasts several minutes (3min on my computer)

src/methods/Chardon_tom/utils.py

+import numpy as np
+import pywt
+
+def get_neighbors (img,channel,i,j,N,M):
+
+    P1 = img[(i-1)%N,(j-1)%M,channel]
+    P2 = img[(i-1)%N,j%M,channel]
+    P3 = img[(i-1)%N,(j+1)%M,channel]
+    P4 = img[i%N,(j-1)%M,channel]
+    P5 = img[i%N,j%M,channel]
+    P6 = img[i%N,(j+1)%M,channel]
+    P7 = img[(i+1)%N,(j-1)%M,channel]
+    P8 = img[(i+1)%N,j%M,channel]
+    P9 = img[(i+1)%N,(j+1)%M,channel]
+
+    return np.array([P1,P2,P3,P4,P5,P6,P7,P8,P9])
+
+
+def get_derivatives(neighbors):
+
+    [P1, P2, P3, P4, P5, P6, P7, P8, P9] = neighbors
+
+    D_x = (P4 - P6)/2
+    D_y = (P2 - P8)/2
+    D_xd = (P3 - P7)/(2*np.sqrt(2))
+    D_yd = (P1 - P9)/(2*np.sqrt(2))
+
+    return ([D_x, D_y, D_xd, D_yd])
+
+
+def get_weights(mosaic_image, i, j, channel, N, M):
+
+    derivatives_neigbors = []
+    for l in range(-1, 2):
+        for L in range(-1, 2):
+            derivatives_neigbors.append(get_derivatives(
+                get_neighbors(mosaic_image, channel, i+l, j+L, N, M)))
+
+    [Dx, Dy, Dxd, Dyd] = derivatives_neigbors[4]
+    E1 = 1/np.sqrt(1 + Dyd**2 + derivatives_neigbors[0][3]**2)
+    E2 = 1/np.sqrt(1 + Dy**2 + derivatives_neigbors[1][1]**2)
+    E3 = 1/np.sqrt(1 + Dxd**2 + derivatives_neigbors[2][2]**2)
+    E4 = 1/np.sqrt(1 + Dx**2 + derivatives_neigbors[3][0]**2)
+    E6 = 1/np.sqrt(1 + Dxd**2 + derivatives_neigbors[5][2]**2)
+    E7 = 1/np.sqrt(1 + Dy**2 + derivatives_neigbors[6][1]**2)
+    E8 = 1/np.sqrt(1 + Dyd**2 + derivatives_neigbors[7][3]**2)
+    E9 = 1/np.sqrt(1 + Dx**2 + derivatives_neigbors[8][0]**2)
+    E = [E1, E2, E3, E4, E6, E7, E8, E9]
+
+    return E    
+
+
+def interpolate_green(weights, neighbors):
+
+    [E1, E2, E3, E4, E6, E7, E8, E9] = weights
+    [P1, P2, P3, P4, P5, P6, P7, P8, P9] = neighbors
+
+    I5 = (E2*P2 + E4*P4 + E6*P6 + E8*P8)/(E2 + E4 + E6 + E8)
+
+    return (I5)
+
+
+def interpolate_red_blue(weights, neighbors, green_neighbors):
+
+    [E1, E2, E3, E4, E6, E7, E8, E9] = weights
+    [P1, P2, P3, P4, P5, P6, P7, P8, P9] = neighbors
+    [G1, G2, G3, G4, G5, G6, G7, G8, G9] = green_neighbors
+
+    I5 = G5*(E1*P1/G1 + E3*P3/G3 + E7*P7/G7 + E9*P9/G9)/(E1 + E3 + E7 + E9)
+
+    return (I5)
+
+
+def correction_green(res,i,j,N,M):
+
+
+    [G1,G2,G3,G4,G5,G6,G7,G8,G9] = get_neighbors(res,1,i,j,N,M)
+    [R1,R2,R3,R4,R5,R6,R7,R8,R9] = get_neighbors(res,0,i,j,N,M)
+    [B1,B2,B3,B4,B5,B6,B7,B8,B9] = get_neighbors(res,2,i,j,N,M)
+    [E1,E2,E3,E4,E6,E7,E8,E9] = get_weights(res,i,j,1,N,M)
+
+    Gb5 = R5*((E2*G2)/B2 + (E4*G4)/B4 + (E6*G6)/B6 + (E8*G8)/B8)/(E2 + E4 + E6 + E8)
+    Gr5 = B5*((E2*G2)/R2 + (E4*G4)/R4 + (E6*G6)/R6 + (E8*G8)/R8)/(E2 + E4 + E6 + E8)
+
+    G5 = (Gb5 + Gr5)/2
+
+    return G5
+
+def correction_red(res,i,j,N,M) :
+
+    [G1,G2,G3,G4,G5,G6,G7,G8,G9] = get_neighbors(res,1,i,j,N,M)
+    [R1,R2,R3,R4,R5,R6,R7,R8,R9] = get_neighbors(res,0,i,j,N,M)
+    [E1,E2,E3,E4,E6,E7,E8,E9] = get_weights(res,i,j,0,N,M)
+
+    R5 = G5*((E1*R1)/G1 + (E2*R2)/G2 + (E3*R3)/G3 + (E4*R4)/G4 + (E6*R6)/G6 + (E7*R7)/G7 + (E8*R8)/G8 + (E9*R9)/G9)/(E1 + E2 + E3 + E4 + E6 + E7 + E8 + E9)
+
+    return R5
+
+def correction_blue(res,i,j,N,M) :
+
+    [G1,G2,G3,G4,G5,G6,G7,G8,G9] = get_neighbors(res,1,i,j,N,M)
+    [B1,B2,B3,B4,B5,B6,B7,B8,B9] = get_neighbors(res,2,i,j,N,M)
+    [E1,E2,E3,E4,E6,E7,E8,E9] = get_weights(res,i,j,2,N,M)
+
+    B5 = G5*((E1*B1)/G1 + (E2*B2)/G2 + (E3*B3)/G3 + (E4*B4)/G4 + (E6*B6)/G6 + (E7*B7)/G7 + (E8*B8)/G8 + (E9*B9)/G9)/(E1 + E2 + E3 + E4 + E6 + E7 + E8 + E9)
+
+    return B5
+
+
+
+

src/methods/GERAYELI_Kourosh/GERAYELI_Kourosh.py

+
+# Importing libraries
+import os
+import colour
+from colour_demosaicing import (
+    demosaicing_CFA_Bayer_bilinear,
+    demosaicing_CFA_Bayer_Malvar2004,
+    demosaicing_CFA_Bayer_Menon2007,
+    mosaicing_CFA_Bayer)
+
+from src.utils import psnr,ssim
+
+# Image path
+image_pathes = ['images/img_1.png','images/img_2.png','images/img_3.png','images/img_4.png']
+
+for i in image_pathes:
+    LIGHTHOUSE_IMAGE = colour.io.read_image(i)
+    img = LIGHTHOUSE_IMAGE
+    colour.plotting.plot_image(
+        colour.cctf_encoding(LIGHTHOUSE_IMAGE))
+
+
+    # Mosaicing
+    CFA = mosaicing_CFA_Bayer(LIGHTHOUSE_IMAGE)
+
+    colour.plotting.plot_image(
+        colour.cctf_encoding(CFA),
+        text_kwargs={'text': 'Lighthouse - CFA - RGGB'})
+
+    colour.plotting.plot_image(
+        colour.cctf_encoding(mosaicing_CFA_Bayer(LIGHTHOUSE_IMAGE, 'BGGR')), 
+        text_kwargs={'text': 'Lighthouse - CFA - BGGR'});
+
+
+    # Demosaicing bilinear
+    colour.plotting.plot_image(
+        colour.cctf_encoding(demosaicing_CFA_Bayer_bilinear(CFA)), 
+        text_kwargs={'text': 'Demosaicing - Bilinear'});
+
+    recons_bilinear = colour.cctf_encoding(demosaicing_CFA_Bayer_bilinear(CFA))
+
+
+    # demosaicing Malvar
+    recons_malvar = colour.cctf_encoding(demosaicing_CFA_Bayer_Malvar2004(CFA))
+    colour.plotting.plot_image(
+        colour.cctf_encoding(demosaicing_CFA_Bayer_Malvar2004(CFA)), 
+        text_kwargs={'text': 'Demosaicing - Malvar (2004)'});
+
+    # demosaicing Menon
+    recons_menon = colour.cctf_encoding(demosaicing_CFA_Bayer_Menon2007(CFA))
+    colour.plotting.plot_image(
+        colour.cctf_encoding(demosaicing_CFA_Bayer_Menon2007(CFA)), 
+        text_kwargs={'text': 'Demosaicing - Menon (2007)'});
+
+
+    print('bilinear : ')
+    print(f'PSNR: {psnr(img, recons_bilinear):.2f}')
+    print(f'SSIM: {ssim(img, recons_bilinear):.4f}')
+
+    print('Malvar : ')
+    print(f'PSNR: {psnr(img, recons_malvar):.2f}')
+    print(f'SSIM: {ssim(img, recons_malvar):.4f}')
+
+    print('Menon')
+    print(f'PSNR: {psnr(img, recons_menon):.2f}')
+    print(f'SSIM: {ssim(img, recons_menon):.4f}')
\ No newline at end of file

src/methods/Samanos_Thomas/Readme.md

+### How to use ?
+There are 2 functions that can be called in reconstruct.py :
+f.cfa_reconstruction_1 that uses the method 1 
+and f.cfa_reconstruction that uses the method 2
+
+They use the same arguments which are y , cfa and input_shape.

src/methods/Samanos_Thomas/function.py

+
+import numpy as np
+import matplotlib.pyplot as plt
+from src.checks import check_cfa, check_rgb 
+from src.forward_model import CFA
+import src.methods.Samanos_Thomas.version1 as v1
+import src.methods.Samanos_Thomas.version2 as v2
+from scipy import ndimage
+
+
+def cfa_reconstruction(y: np.ndarray, cfa:str, input_shape: tuple) -> np.ndarray:
+    """Performs demosaicking on y.
+
+    Args:
+        cfa (str): Name of the CFA. Can be bayer or quad_bayer.
+
+    Returns:
+        np.ndarray: Demosaicked image.
+    """
+    if cfa == "bayer":
+        op = v2.bayer_reconstruction(y,input_shape)
+    elif cfa == "quad_bayer":
+        op = v2.quad_bayer_reconstruction(y,input_shape)
+    else:
+        op = np.zeros(input_shape)
+    return op
+
+def cfa_reconstruction_1(y: np.ndarray, cfa:str, input_shape: tuple) -> np.ndarray:
+    """Performs demosaicking on y.
+
+    Args:
+        cfa (str): Name of the CFA. Can be bayer or quad_bayer.
+
+    Returns:
+        np.ndarray: Demosaicked image.
+    """
+    if cfa == "bayer":
+        op = v1.bayer_reconstruction(y,input_shape)
+    elif cfa == "quad_bayer":
+        op = v1.quad_bayer_reconstruction(y,input_shape)
+    else:
+        op = np.zeros(input_shape)
+    return op
\ No newline at end of file

src/methods/Samanos_Thomas/reconstruct.py

+"""The main file for the reconstruction.
+This file should NOT be modified except the body of the 'run_reconstruction' function.
+Students can call their functions (declared in others files of src/methods/your_name).
+"""
+
+
+import numpy as np
+
+import src.methods.Samanos_Thomas.function as f
+from src.forward_model import CFA
+
+def run_reconstruction(y: np.ndarray, cfa: str) -> np.ndarray:
+    """Performs demosaicking on y.
+
+    Args:
+        y (np.ndarray): Mosaicked image to be reconstructed.
+        cfa (str): Name of the CFA. Can be bayer or quad_bayer.
+
+    Returns:
+        np.ndarray: Demosaicked image.
+    """
+
+    input_shape = (y.shape[0], y.shape[1], 3)
+    Y = f.cfa_reconstruction(y, cfa,input_shape)
+    return Y
+
+####
+####
+####
+
+####      ####                ####        #############
+####      ######              ####      ##################
+####      ########            ####      ####################
+####      ##########          ####      ####        ########
+####      ############        ####      ####            ####
+####      ####  ########      ####      ####            ####
+####      ####    ########    ####      ####            ####
+####      ####      ########  ####      ####            ####
+####      ####  ##    ######  ####      ####          ######
+####      ####  ####      ##  ####      ####    ############
+####      ####  ######        ####      ####    ##########
+####      ####  ##########    ####      ####    ########
+####      ####      ########  ####      ####
+####      ####        ############      ####
+####      ####          ##########      ####
+####      ####            ########      ####
+####      ####              ######      ####
+
+# 2023
+# Authors: Mauro Dalla Mura and Matthieu Muller

src/methods/Samanos_Thomas/version1.py

+import numpy as np
+from skimage.transform import resize
+
+def upscale(img: np.ndarray, shape: tuple, pas: int) -> np.ndarray:
+    """Upscales the acquisition to the shape of the desired reconstruction.
+
+    Args:
+        img (np.ndarray): mosaique image in smaller size.
+        shape (tuple): Shape of the output.
+
+    Returns:
+        np.ndarray: Upscaled image.
+    """
+    return resize(img, shape, anti_aliasing=True)
+
+def bayer_reconstruction(y : np.ndarray, input_shape : tuple) -> np.ndarray:
+    op_shape = (input_shape[0]//2,input_shape[1]//2,3)
+    op = np.zeros(op_shape)
+    conv = np.zeros((2,2,3))
+    liste = [(0,1,0),(0,0,1),(1,1,1),(1,0,2)]
+    for el in liste :
+        conv[el]=1
+        if el[2]==1:
+            conv[el]=1/2
+    for i in range (3):
+        op[:,:,i] = conv2d(y,conv[:,:,i], 2)
+    op1 = upscale(op,(input_shape[0],input_shape[1],3),2)
+    return op1
+
+def quad_bayer_reconstruction(y : np.ndarray, input_shape : tuple) -> np.ndarray:
+    op_shape = (input_shape[0]//4,input_shape[1]//4,3)
+    op = np.zeros(op_shape)
+    conv = np.zeros((4,4,3))
+    liste = [(2,0,2),(3,0,2),(2,1,2),(3,1,2),(0,0,1),(1,0,1),(0,1,1),(1,1,1),(2,2,1),(2,3,1),(3,2,1),(3,3,1),(0,2,0),(0,3,0),(1,2,0),(1,3,0)]
+    for el in liste :
+        conv[el]=1/4
+        if el[2]==1:
+            conv[el]=1/8
+    for i in range (3):
+        op[:,:,i] = conv2d(y,conv[:,:,i], 4)
+    op1 = upscale(op,(input_shape[0],input_shape[1],3),4)
+    return op1
+
+def conv2d(img : np.ndarray, conv : np.ndarray, pas :int):
+    y_shape = (img.shape[0], img.shape[1])
+    op = np.zeros((y_shape[0]//pas,y_shape[1]//pas))
+    for i in range(0,y_shape[0],pas):
+        for j in range(0,y_shape[1],pas):
+            op[i//pas,j//pas] = np.sum(img[i:i+pas,j:j+pas]*conv)
+    return op

src/methods/Samanos_Thomas/version2.py

+import numpy as np
+from scipy.signal import convolve2d
+
+CONV2 = np.array([[1/4,1/2,1/4],[1/2,1,1/2],[1/4,1/2,1/4]])
+
+def bayer_reconstruction(y : np.ndarray, input_shape : tuple) -> np.ndarray:
+    op_shape = (input_shape[0],input_shape[1],3)
+    op = np.zeros(op_shape)
+    conv = np.zeros((2,2,3))
+    liste = [(0,1,0),(0,0,1),(1,1,1),(1,0,2)]
+    for el in liste :
+        conv[el]=1
+    op = masq(y,conv, 2)
+    op1 = np.zeros_like(op)
+    for color in range(3):
+        op1[:,:,color] = convolve2d(op[:,:,color],CONV2,mode ="same")
+        if color == 1 :
+            op1[:,:,color] = op1[:,:,color]/2
+    return op1
+
+CONV4 = np.array([[11/156,43/312,2/13,43/312,11/156],
+                 [43/312,1/6,1/4,1/6,43/312],
+                 [2/13,1/4,1/3,1/4,2/13],
+                 [43/312,1/6,1/4,1/6,43/312],
+                 [11/156,43/312,2/13,43/312,11/156]])
+
+def quad_bayer_reconstruction(y : np.ndarray, input_shape : tuple) -> np.ndarray:
+    op_shape = (input_shape[0],input_shape[1],3)
+    op = np.zeros(op_shape)
+    conv = np.zeros((4,4,3))
+    liste = [(2,0,2),(3,0,2),(2,1,2),(3,1,2),(0,0,1),(1,0,1),(0,1,1),(1,1,1),(2,2,1),(2,3,1),(3,2,1),(3,3,1),(0,2,0),(0,3,0),(1,2,0),(1,3,0)]
+    for el in liste :
+        conv[el]=1
+    op = masq(y,conv, 4)
+    op1 = np.zeros_like(op)
+    for color in range(3):
+        op1[:,:,color] = convolve2d(op[:,:,color],CONV4,mode ="same")
+        if color == 1 :
+            op1[:,:,color] = op1[:,:,color]/2
+    return op1
+
+def masq(img : np.ndarray,conv:np.ndarray, pas:int) ->np.ndarray:
+    masq = np.zeros((img.shape[0],img.shape[1],3))
+    po = np.zeros((img.shape[0],img.shape[1],3))
+    for color in range (3):
+        for i in range(0,po.shape[0],pas):
+            for j in range(0,po.shape[1],pas):
+                masq[i:i+pas,j:j+pas,color] = conv[:,:,color]
+        po[:,:,color] = masq[:,:,color] * img
+    return po
\ No newline at end of file

src/methods/Samia_Bouddahab/function.py

+# -*- coding: utf-8 -*-
+"""
+Created on Mon Feb 12 07:38:51 2024
+
+@author: etudiant
+"""
+import numpy as np
+from scipy.signal import convolve2d
+from src.forward_model import CFA
+from scipy.ndimage import convolve
+
+
+def malvar_demosaicing(op: CFA, y: np.ndarray) -> np.ndarray:
+    """
+    Performs a malvar interpolation.
+
+    Args:
+        op (CFA): CFA operator.
+        y (np.ndarray): Mosaicked image.
+
+    Returns:
+        np.ndarray: Demosaicked image.
+    """
+    z = op.adjoint(y)
+    # Definnig Masks
+    R_m = op.mask[:, :, 0]
+    G_m = op.mask[:, :, 1]
+    B_m = op.mask[:, :, 2]
+
+    GR_GB = np.array([
+        [0.0, 0.0, -1.0, 0.0, 0.0],
+        [0.0, 0.0, 2.0, 0.0, 0.0],
+        [-1.0, 2.0, 4.0, 2.0, -1.0],
+        [0.0, 0.0, 2.0, 0.0, 0.0],
+        [0.0, 0.0, -1.0, 0.0, 0.0],
+    ]) / 8
+
+    Rg_RB_Bg_BR = np.array(
+        [
+            [0.0, 0.0, 0.5, 0.0, 0.0],
+            [0.0, -1.0, 0.0, -1.0, 0.0],
+            [-1.0, 4.0, 5.0, 4.0, -1.0],
+            [0.0, -1.0, 0.0, -1.0, 0.0],
+            [0.0, 0.0, 0.5, 0.0, 0.0],
+        ]) / 8
+
+    Rg_BR_Bg_RB = np.transpose(Rg_RB_Bg_BR)
+
+    Rb_BB_Br_RR = np.array(
+        [
+            [0.0, 0.0, -1.5, 0.0, 0.0],
+            [0.0, 2.0, 0.0, 2.0, 0.0],
+            [-1.5, 0.0, 6.0, 0.0, -1.5],
+            [0.0, 2.0, 0.0, 2.0, 0.0],
+            [0.0, 0.0, -1.5, 0.0, 0.0],
+        ]) / 8
+
+    R = y * R_m
+    G = y * G_m
+    B = y * B_m
+
+    del G_m
+
+    G = np.where(np.logical_or(R_m == 1, B_m == 1), convolve(y, GR_GB), G)
+
+    RBg_RBBR = convolve(y, Rg_RB_Bg_BR)
+    RBg_BRRB = convolve(y, Rg_BR_Bg_RB)
+    RBgr_BBRR = convolve(y, Rb_BB_Br_RR)
+
+    del GR_GB, Rg_RB_Bg_BR, Rg_BR_Bg_RB, Rb_BB_Br_RR
+
+    # Red rows.
+    R_r = np.transpose(np.any(R_m == 1, axis=1)[None]) * np.ones(R.shape)
+    # Red columns.
+    R_c = np.any(R_m == 1, axis=0)[None] * np.ones(R.shape)
+    # Blue rows.
+    B_r = np.transpose(np.any(B_m == 1, axis=1)[None]) * np.ones(B.shape)
+    # Blue columns
+    B_c = np.any(B_m == 1, axis=0)[None] * np.ones(B.shape)
+
+    del R_m, B_m
+
+    R = np.where(np.logical_and(R_r == 1, B_c == 1), RBg_RBBR, R)
+    R = np.where(np.logical_and(B_r == 1, R_c == 1), RBg_BRRB, R)
+
+    B = np.where(np.logical_and(B_r == 1, R_c == 1), RBg_RBBR, B)
+    B = np.where(np.logical_and(R_r == 1, B_c == 1), RBg_BRRB, B)
+
+    R = np.where(np.logical_and(B_r == 1, B_c == 1), RBgr_BBRR, R)
+    B = np.where(np.logical_and(R_r == 1, R_c == 1), RBgr_BBRR, B)
+
+    del RBg_RBBR, RBg_BRRB, RBgr_BBRR, R_r, R_c, B_r, B_c
+    # Combine channels
+    return np.clip(np.stack((R, G, B), axis=-1), 0, 1)

src/methods/Samia_Bouddahab/reconstruct.py

+"""The main file for the reconstruction.
+This file should NOT be modified except the body of the 'run_reconstruction' function.
+Students can call their functions (declared in others files of src/methods/your_name).
+"""
+
+
+import numpy as np
+
+from src.forward_model import CFA
+from src.methods.Samia_Bouddahab.function import malvar_demosaicing
+
+
+def run_reconstruction(y: np.ndarray, cfa: str) -> np.ndarray:
+    """Performs demosaicking on y.
+
+    Args:
+        y (np.ndarray): Mosaicked image to be reconstructed.
+        cfa (str): Name of the CFA. Can be bayer or quad_bayer.
+
+    Returns:
+        np.ndarray: Demosaicked image.
+    """
+    # Performing the reconstruction.
+    # TODO
+    input_shape = (y.shape[0], y.shape[1], 3)
+    op = CFA(cfa, input_shape)
+    res = malvar_demosaicing(op, y)
+    return np.zeros(op.input_shape)
+
+
+####
+####
+####
+
+####      ####                ####        #############
+####      ######              ####      ##################
+####      ########            ####      ####################
+####      ##########          ####      ####        ########
+####      ############        ####      ####            ####
+####      ####  ########      ####      ####            ####
+####      ####    ########    ####      ####            ####
+####      ####      ########  ####      ####            ####
+####      ####  ##    ######  ####      ####          ######
+####      ####  ####      ##  ####      ####    ############
+####      ####  ######        ####      ####    ##########
+####      ####  ##########    ####      ####    ########
+####      ####      ########  ####      ####
+####      ####        ############      ####
+####      ####          ##########      ####
+####      ####            ########      ####
+####      ####              ######      ####
+
+# 2023
+# Authors: Mauro Dalla Mura and Matthieu Muller

src/methods/TheoLafond/reconstruct.py

+"""The main file for the reconstruction.
+This file should NOT be modified except the body of the 'run_reconstruction' function.
+Students can call their functions (declared in others files of src/methods/your_name).
+"""
+
+
+import numpy as np
+from scipy.ndimage import generic_filter, convolve
+import warnings
+from tqdm import tqdm
+from src.forward_model import CFA
+
+
+
+def regression_filter_func(chans):
+    """computes the linear regression coefficients between channel :
+    red and green
+    blue and green
+    green and red
+    green and blue
+
+
+    Returns:
+        a is high order coef and b is 0 order.
+        _type_: a_red, b_red, a_blue, b_blue, a_green_red, b_green_red, a_green_blue, b_green_blue
+    """
+    red = chans[:,:,0].flatten()
+    green = chans[:,:,1].flatten()
+    blue = chans[:,:,2].flatten()
+    nonzeros_red = np.nonzero(red)
+    nonzeros_blue = np.nonzero(blue)
+    return np.concatenate((np.polyfit(red[nonzeros_red], green[nonzeros_red], deg=1), np.polyfit(blue[nonzeros_blue], green[nonzeros_blue], deg=1), np.polyfit(green[nonzeros_red], red[nonzeros_red], deg=1), np.polyfit(green[nonzeros_blue], blue[nonzeros_blue], deg=1)), axis=0)
+
+def personal_generic_filter(a, func, footprint, output, msg=""):
+    """Apply func on each footprint of a.
+    """
+
+    # Suppress RankWarning
+    with warnings.catch_warnings():
+        warnings.simplefilter('ignore', np.RankWarning)
+        for i in tqdm(range(a.shape[0]), desc=msg):
+            for j in range(a.shape[1]):
+                fen = np.ones((footprint.shape[0], footprint.shape[1], a.shape[2]))
+                if i+footprint.shape[0] > a.shape[0]:
+                    i_end = a.shape[0]
+                    i_start = i_end - footprint.shape[0]
+                else:
+                    i_start = i
+                    i_end = i + footprint.shape[0]
+                if j+footprint.shape[1] > a.shape[1]:
+                    j_end = a.shape[1]
+                    j_start = j_end - footprint.shape[1]
+                else:
+                    j_start = j
+                    j_end = j + footprint.shape[1]
+                fen[:i_end-i_start, :j_end-j_start] = a[i_start:i_end, j_start:j_end]
+                output[i, j, :] = func(a[i_start:i_end, j_start:j_end])
+
+def combine_directions(bh,bv):
+    combo = np.zeros(bh.shape)
+    combo[:,:,0] = bh[:,:,0] + bv[:,:,0]
+    combo[:,:,0][(bh[:,:,0]!=0) * (bv[:,:,0]!=0)] = (bh[:,:,0][(bh[:,:,0]!=0) * (bv[:,:,0]!=0)] + bv[:,:,0][(bh[:,:,0]!=0) * (bv[:,:,0]!=0)])/2
+    combo[:,:,0][(bh[:,:,0]==0) * (bv[:,:,0]==0)] = 0
+
+    combo[:,:,1] = bh[:,:,1]/2 + bv[:,:,1]/2
+    combo[:,:,2] = bh[:,:,2] + bv[:,:,2]
+    combo[:,:,2][(bh[:,:,2]!=0) * (bv[:,:,2]!=0)] = (bh[:,:,2][(bh[:,:,2]!=0) * (bv[:,:,2]!=0)] + bv[:,:,2][(bh[:,:,2]!=0) * (bv[:,:,2]!=0)])/2
+    combo[:,:,2][(bh[:,:,2]==0) * (bv[:,:,2]==0)] = 0
+    
+
+    for i in range(combo.shape[0]):
+        for j in range(combo.shape[1]):
+            moy = []
+            if i != 0 :
+                moy.append(combo[i-1,j,:])
+            if i != combo.shape[0]-1 :
+                moy.append(combo[i+1,j,:])
+            if j != 0 :
+                moy.append(combo[i,j-1,:])
+            if j != combo.shape[1]-1 :
+                moy.append(combo[i,j+1,:])
+            moy = np.stack(moy).mean(axis=0)
+            if combo[i,j,0] == 0:
+                combo[i,j,0] = moy[0]
+            if combo[i,j,2] == 0:
+                combo[i,j,2] = moy[2]
+    return combo
+
+
+def run_reconstruction(y: np.ndarray, cfa: str) -> np.ndarray:
+    """Performs demosaicking on y.
+
+    Args:
+        y (np.ndarray): Mosaicked image to be reconstructed.
+        cfa (str): Name of the CFA. Can be bayer or quad_bayer.
+
+    Returns:
+        np.ndarray: Demosaicked image.
+    """
+    # Performing the reconstruction.
+
+    input_shape = (y.shape[0], y.shape[1], 3)
+    op = CFA(cfa, input_shape)
+    adjoint = op.adjoint(y)
+
+    bh = np.zeros(adjoint.shape)
+    # Horizontal interpolation
+    for i in range(3):
+        for j in range(adjoint.shape[0]):
+            non_zero = np.nonzero(adjoint[j, :, i])[0]
+            if len(non_zero) == 0:
+                continue
+            bh[j, :, i] = np.interp(np.arange(adjoint.shape[1]), non_zero, adjoint[j, :, i][non_zero])
+
+    bv = np.zeros(adjoint.shape)
+    # Vertical interpolation
+    for i in range(3):
+        for j in range(adjoint.shape[1]):
+            non_zero = np.nonzero(adjoint[:, j, i])[0]
+            if len(non_zero) == 0:
+                continue
+            bv[:, j, i] = np.interp(np.arange(adjoint.shape[0]), non_zero, adjoint[:, j, i][non_zero])
+
+    # Residuals interpolation
+    M = 3
+    N = 5
+    kernel_hor = np.ones((M, N))/(M*N)
+    kernel_ver = np.ones((N, M))/(M*N)
+
+    # Horizontal Regression filtering
+    regh = np.zeros((bh.shape[0], bh.shape[1], 8))
+    personal_generic_filter(bh, regression_filter_func, kernel_hor, regh,msg="Regression filtering horizontal")
+    ch = np.zeros(bh.shape)
+    for i in range(regh.shape[-1]):
+        regh[:, :, i] = convolve(regh[:, :, i], kernel_hor)
+    ch[:,:,0] = regh[:,:,0]*bh[:,:,0] + regh[:,:,1]
+    ch[:,:,1] = (regh[:,:,4]*bh[:,:,1] + regh[:,:,5] + regh[:,:,6]*bh[:,:,1] + regh[:,:,7])/2
+    ch[:,:,2] = regh[:,:,2]*bh[:,:,2] + regh[:,:,3]
+    # return ch[:,:,1]
+    dh = ch - adjoint
+    dh[adjoint==0] = 0
+    # interpolation
+    for i in range(3):
+        for j in range(adjoint.shape[0]):
+            non_zero = np.nonzero(dh[j, :, i])[0]
+            if len(non_zero) == 0:
+                continue
+            ch[j, :, i] -= np.interp(np.arange(adjoint.shape[1]), non_zero, dh[j, :, i][non_zero])
+    ch[bh == 0] = 0
+    bh = ch.copy()
+    # return bh[:,:,0]
+
+    # Vertical Regression filtering
+    regv = np.zeros((bv.shape[0], bv.shape[1], 8))
+    personal_generic_filter(bv, regression_filter_func, kernel_ver, regv,msg="Regression filtering vertical")
+    cv = np.zeros(bv.shape)
+    for i in range(regv.shape[-1]):
+        regv[:, :, i] = convolve(regv[:, :, i], kernel_ver)
+    cv[:,:,0] = regv[:,:,0]*bv[:,:,0] + regv[:,:,1]
+    cv[:,:,1] = (regv[:,:,4]*bv[:,:,1] + regv[:,:,5] + regv[:,:,6]*bv[:,:,1] + regv[:,:,7])/2
+    cv[:,:,2] = regv[:,:,2]*bv[:,:,2] + regv[:,:,3]
+    dv = cv - adjoint
+    dv[adjoint==0] = 0
+    # interpolation
+    for i in range(3):
+        for j in range(adjoint.shape[1]):
+            non_zero = np.nonzero(dv[:, j, i])[0]
+            if len(non_zero) == 0:
+                continue
+            cv[:, j, i] -= np.interp(np.arange(adjoint.shape[0]), non_zero, dv[:, j, i][non_zero])
+    cv[bv == 0] = 0
+    bv = cv.copy()
+
+    return combine_directions(bh,bv)
+
+    
+
+
+if __name__ == "__main__":
+    from src.utils import load_image, save_image, psnr, ssim
+    from src.forward_model import CFA
+    image_path = 'images/img_4.png'
+
+    img = load_image(image_path)
+
+    cfa_name = 'bayer' # bayer or quad_bayer
+    op = CFA(cfa_name, img.shape)
+    y = op.direct(img)
+    res = run_reconstruction(y, cfa_name)
+    print('PSNR:', psnr(img, res))
+    print('SSIM:', ssim(img, res))
+    save_image('output/bh.png', res)
+
+
+
+####
+####
+####
+
+####      ####                ####        #############
+####      ######              ####      ##################
+####      ########            ####      ####################
+####      ##########          ####      ####        ########
+####      ############        ####      ####            ####
+####      ####  ########      ####      ####            ####
+####      ####    ########    ####      ####            ####
+####      ####      ########  ####      ####            ####
+####      ####  ##    ######  ####      ####          ######
+####      ####  ####      ##  ####      ####    ############
+####      ####  ######        ####      ####    ##########
+####      ####  ##########    ####      ####    ########
+####      ####      ########  ####      ####
+####      ####        ############      ####
+####      ####          ##########      ####
+####      ####            ########      ####
+####      ####              ######      ####
+
+# 2023
+# Authors: Mauro Dalla Mura and Matthieu Muller