add

Author: yoyoyo-yo

Question_81_90/README.md

@@ -167,7 +167,7 @@ Q.85では最も色が近い画像が予測クラスであるとしたが、そ
 
 これを回避するために、ここでは色合いが近い画像を3つ選び、それらの多数決によって予測クラスを決定し、Accuracyを計算せよ。
 
-このように特徴が近いものを学習データから3つ選んで判断する手法をk近傍(k-NN: k-Nearest Neighbor)という。Q.85.のNN法はk=1の場合とみれる。
+このように特徴が近いものを学習データからk個選んで判断する手法をk近傍(k-NN: k-Nearest Neighbor)という。Q.85.のNN法はk=1の場合とみれる。
 
 
 答え

Question_81_90/answers/answer_81.jpg

@@ -2,68 +2,94 @@
 import numpy as np
 import matplotlib.pyplot as plt
 
-# Read image
-img = cv2.imread("thorino.jpg").astype(np.float32)
-H, W, C = img.shape
+# Hessian corner detection
+def Hessian_corner(img):
+
+	## Grayscale
+	def BGR2GRAY(img):
+		gray = 0.2126 * img[..., 2] + 0.7152 * img[..., 1] + 0.0722 * img[..., 0]
+		gray = gray.astype(np.uint8)
+		return gray
+
+	## Sobel
+	def Sobel_filtering(gray):
+		# get shape
+		H, W = gray.shape
+
+		# sobel kernel
+		sobely = np.array(((1, 2, 1),
+						(0, 0, 0),
+						(-1, -2, -1)), dtype=np.float32)
+
+		sobelx = np.array(((1, 0, -1),
+						(2, 0, -2),
+						(1, 0, -1)), dtype=np.float32)
 
-## Grayscale
-gray = 0.2126 * img[..., 2] + 0.7152 * img[..., 1] + 0.0722 * img[..., 0]
-gray = gray.astype(np.uint8)
+		# padding
+		tmp = np.pad(gray, (1, 1), 'edge')
 
-## Sobel
-sobely = np.array(((1, 2, 1),
-                   (0, 0, 0),
-                   (-1, -2, -1)), dtype=np.float32)
+		# prepare
+		Ix = np.zeros_like(gray, dtype=np.float32)
+		Iy = np.zeros_like(gray, dtype=np.float32)
 
-sobelx = np.array(((1, 0, -1),
-                   (2, 0, -2),
-                   (1, 0, -1)), dtype=np.float32)
+		# get differential
+		for y in range(H):
+			for x in range(W):
+				Ix[y, x] = np.mean(tmp[y : y  + 3, x : x + 3] * sobelx)
+				Iy[y, x] = np.mean(tmp[y : y + 3, x : x + 3] * sobely)
+			
+		Ix2 = Ix ** 2
+		Iy2 = Iy ** 2
+		Ixy = Ix * Iy
 
-tmp = np.pad(gray, (1, 1), 'edge')
+		return Ix2, Iy2, Ixy
 
-Ix = np.zeros_like(gray, dtype=np.float32)
-Iy = np.zeros_like(gray, dtype=np.float32)
+		
 
-for y in range(H):
-    for x in range(W):
-        Ix[y, x] = np.mean(tmp[y:y+3, x:x+3] * sobelx)
-        Iy[y, x] = np.mean(tmp[y:y+3, x:x+3] * sobely)
-     
-tmp = np.pad(Ix, (1, 1), 'edge')
+	## Hessian
+	def corner_detect(gray, Ix2, Iy2, Ixy):
+		# get shape
+		H, W = gray.shape
 
-Ix2 = np.zeros_like(gray, dtype=np.float32)
-IxIy = np.zeros_like(gray, dtype=np.float32)
+		# prepare for show detection
+		out = np.array((gray, gray, gray))
+		out = np.transpose(out, (1,2,0))
 
-for y in range(H):
-    for x in range(W):
-        Ix2[y, x] = np.mean(tmp[y:y+3, x:x+3] * sobelx)
-        IxIy[y, x] = np.mean(tmp[y:y+3, x:x+3] * sobely)
+		# get Hessian value
+		Hes = np.zeros((H, W))
 
-tmp = np.pad(Iy, (1, 1), 'edge')
+		for y in range(H):
+			for x in range(W):
+				Hes[y,x] = Ix2[y,x] * Iy2[y,x] - Ixy[y,x] ** 2
 
-Iy2 = np.zeros_like(gray, dtype=np.float32)
+		## Detect Corner and show
+		for y in range(H):
+			for x in range(W):
+				if Hes[y,x] == np.max(Hes[max(y-1, 0) : min(y+2, H), max(x-1, 0) : min(x+2, W)]) and Hes[y, x] > np.max(Hes) * 0.1:
+					out[y, x] = [0, 0, 255]
 
-for y in range(H):
-    for x in range(W):
-        Iy2[y, x] = np.mean(tmp[y:y+3, x:x+3] * sobely)
+		out = out.astype(np.uint8)
 
-out = np.array((gray, gray, gray))
-out = np.transpose(out, (1,2,0))
+		return out
 
-## Hessian
-Hes = np.zeros((H, W))
+	
+	# 1. grayscale
+	gray = BGR2GRAY(img)
 
-for y in range(H):
-    for x in range(W):
-        Hes[y,x] = Ix2[y,x] * Iy2[y,x] - IxIy[y,x] ** 2
+	# 2. get difference image
+	Ix2, Iy2, Ixy = Sobel_filtering(gray)
 
-## Detect Corner
-for y in range(H):
-    for x in range(W):
-        if Hes[y,x] == np.max(Hes[max(y-1,0):min(y+2,H), max(x-1,0):min(x+2,W)]) and Hes[y,x] > np.max(Hes)*0.1:
-            out[y, x] = [0, 0, 255]
+	# 3. corner detection
+	out = corner_detect(gray, Ix2, Iy2, Ixy)
+
+	return out
+
+
+# Read image
+img = cv2.imread("thorino.jpg").astype(np.float32)
 
-out = out.astype(np.uint8)
+# Hessian corner detection
+out = Hessian_corner(img)
 
 cv2.imwrite("out.jpg", out)
 cv2.imshow("result", out)

Question_81_90/answers/answer_81.py

@@ -2,79 +2,111 @@
 import numpy as np
 import matplotlib.pyplot as plt
 
+
+# Harris corner detection
+def Harris_corner_step1(img):
+
+	## Grayscale
+	def BGR2GRAY(img):
+		gray = 0.2126 * img[..., 2] + 0.7152 * img[..., 1] + 0.0722 * img[..., 0]
+		gray = gray.astype(np.uint8)
+		return gray
+
+	## Sobel
+	def Sobel_filtering(gray):
+		# get shape
+		H, W = gray.shape
+
+		# sobel kernel
+		sobely = np.array(((1, 2, 1),
+						(0, 0, 0),
+						(-1, -2, -1)), dtype=np.float32)
+
+		sobelx = np.array(((1, 0, -1),
+						(2, 0, -2),
+						(1, 0, -1)), dtype=np.float32)
+
+		# padding
+		tmp = np.pad(gray, (1, 1), 'edge')
+
+		# prepare
+		Ix = np.zeros_like(gray, dtype=np.float32)
+		Iy = np.zeros_like(gray, dtype=np.float32)
+
+		# get differential
+		for y in range(H):
+			for x in range(W):
+				Ix[y, x] = np.mean(tmp[y : y  + 3, x : x + 3] * sobelx)
+				Iy[y, x] = np.mean(tmp[y : y + 3, x : x + 3] * sobely)
+			
+		Ix2 = Ix ** 2
+		Iy2 = Iy ** 2
+		Ixy = Ix * Iy
+
+		return Ix2, Iy2, Ixy
+
+
+	# gaussian filtering
+	def gaussian_filtering(I, K_size=3, sigma=3):
+		# get shape
+		H, W = I.shape
+
+		## gaussian
+		I_t = np.pad(I, (K_size // 2, K_size // 2), 'edge')
+
+		# gaussian kernel
+		K = np.zeros((K_size, K_size), dtype=np.float)
+		for x in range(K_size):
+			for y in range(K_size):
+				_x = x - K_size // 2
+				_y = y - K_size // 2
+				K[y, x] = np.exp( -(_x ** 2 + _y ** 2) / (2 * (sigma ** 2)))
+		K /= (sigma * np.sqrt(2 * np.pi))
+		K /= K.sum()
+
+		# filtering
+		for y in range(H):
+			for x in range(W):
+				I[y,x] = np.sum(I_t[y : y + K_size, x : x + K_size] * K)
+				
+		return I
+
+	
+	# 1. grayscale
+	gray = BGR2GRAY(img)
+
+	# 2. get difference image
+	Ix2, Iy2, Ixy = Sobel_filtering(gray)
+
+	# 3. gaussian filtering
+	Ix2 = gaussian_filtering(Ix2, K_size=3, sigma=3)
+	Iy2 = gaussian_filtering(Iy2, K_size=3, sigma=3)
+	Ixy = gaussian_filtering(Ixy, K_size=3, sigma=3)
+
+	# show result
+	plt.subplots_adjust(left=0, right=1, top=1, bottom=0, hspace=0, wspace=0.2)
+
+	plt.subplot(1,3,1)
+	plt.imshow(Ix2, cmap='gray')
+	plt.title("Ix^2")
+	plt.axis("off")
+
+	plt.subplot(1,3,2)
+	plt.imshow(Iy2, cmap='gray')
+	plt.title("Iy^2")
+	plt.axis("off")
+
+	plt.subplot(1,3,3)
+	plt.imshow(Ixy, cmap='gray')
+	plt.title("Ixy")
+	plt.axis("off")
+
+	plt.savefig("out.png")
+	plt.show()
+
+
 # Read image
 img = cv2.imread("thorino.jpg").astype(np.float32)
-H, W, C = img.shape
-
-## Grayscale
-gray = 0.2126 * img[..., 2] + 0.7152 * img[..., 1] + 0.0722 * img[..., 0]
-
-# Harris
-
-## Sobel
-sobely = np.array(((1, 2, 1),
-                   (0, 0, 0),
-                   (-1, -2, -1)), dtype=np.float32)
-
-sobelx = np.array(((1, 0, -1),
-                   (2, 0, -2),
-                   (1, 0, -1)), dtype=np.float32)
-
-tmp = np.pad(gray, (1, 1), 'edge')
-
-Ix = np.zeros_like(gray, dtype=np.float32)
-Iy = np.zeros_like(gray, dtype=np.float32)
-
-for y in range(H):
-    for x in range(W):
-        Ix[y, x] = np.sum(tmp[y:y+3, x:x+3] * sobelx)
-        Iy[y, x] = np.sum(tmp[y:y+3, x:x+3] * sobely)
-
-Ix2 = Ix ** 2
-Iy2 = Iy ** 2
-Ixy = Ix * Iy
-
-## gaussian
-K_size = 3
-sigma = 3
-Ix2_t = np.pad(Ix2, (K_size // 2, K_size // 2), 'edge')
-Iy2_t = np.pad(Iy2, (K_size // 2, K_size // 2), 'edge')
-Ixy_t = np.pad(Ixy, (K_size // 2, K_size // 2), 'edge')
-
-K = np.zeros((K_size, K_size), dtype=np.float)
-for x in range(K_size):
-    for y in range(K_size):
-        _x = x - K_size // 2
-        _y = y - K_size // 2
-        K[y, x] = np.exp( -(_x**2 + _y**2) / (2 * (sigma**2)))
-K /= (sigma * np.sqrt(2 * np.pi))
-K /= K.sum()
-
-for y in range(H):
-    for x in range(W):
-        Ix2[y,x] = np.sum(Ix2_t[y:y+K_size, x:x+K_size] * K)
-        Iy2[y,x] = np.sum(Iy2_t[y:y+K_size, x:x+K_size] * K)
-        Ixy[y,x] = np.sum(Ixy_t[y:y+K_size, x:x+K_size] * K)
-
-out = np.array((gray, gray, gray))
-out = np.transpose(out, (1,2,0))
-
-plt.subplots_adjust(left=0, right=1, top=1, bottom=0, hspace=0, wspace=0.2)
-
-plt.subplot(1,3,1)
-plt.imshow(Ix2, cmap='gray')
-plt.title("Ix^2")
-plt.axis("off")
-
-plt.subplot(1,3,2)
-plt.imshow(Iy2, cmap='gray')
-plt.title("Iy^2")
-plt.axis("off")
-
-plt.subplot(1,3,3)
-plt.imshow(Ixy, cmap='gray')
-plt.title("Ixy")
-plt.axis("off")
-
-plt.savefig("out.png")
-plt.show()
+
+# Harris corner detection step1
+out = Harris_corner_step1(img)