2021 lab

mediaic · Jul 31, 2021 · 282fcee · 282fcee
1 parent 22f89f7
commit 282fcee
Show file tree

Hide file tree

Showing 3 changed files with 280 additions and 21 deletions.
diff --git a/ImageRead.ipynb b/ImageRead.ipynb
diff --git a/Lab_1/Readme.md b/Lab_1/Readme.md
@@ -1,6 +1,5 @@
 # Lab_1 
 **What you will learn :**
- - Image Operation
  - Image Filtering
  - Image PCA & Classification 
 
@@ -23,29 +22,12 @@ You can only use the python packages below:
  - sklearn
  - math
 
-## 1. Image Operation
-
-1.	Load "**fig.jpg**" and try to plot it.
-2.	Resize the image (**fig.jpg**) with scaling factor = 0.25 and try different interpolation methods.
-
-	- cv2.INTER_NEAREST&nbsp;&nbsp;&nbsp;  [fig_1.jpg] 
-	- cv2.INTER_LINEAR     &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;  [fig_2.jpg] 
-	- cv2.INTER_CUBIC     &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;  [fig_3.jpg]
-3. Resize the second image (**fig_2.jpg**) with scaling factor = 4 and try different interpolation methods.
-
-	- cv2.INTER_NEAREST&nbsp;&nbsp;&nbsp;  [fig_4.jpg] 
-	- cv2.INTER_LINEAR     &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;  [fig_5.jpg] 
-	- cv2.INTER_CUBIC     &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;  [fig_6.jpg]
-4.  Convert the color space to *YCbCr* for the upsampled images and the <u>original image</u>, and compute **PSNR** values for "**Y**" channel to the original image (**fig.jpg**). (FLAG: `cv2.COLOR_BGR2YCR_CB`)
-
-**Q**: Which upsampling method performs better?
-
-## 2. Image Filtering
+## 1. Image Filtering
 1.	Load "**fig.jpg**" with **gray** scale.
 2.	Compute and save the results of gradient generated by horizontal Sobel filter and vertical Sobel filter as **fig_grad_x.jpg** and **fig_grad_y.jpg** (with cv2.filter2D).
 3.	Compute and save the gradient magnitude as **fig_grad_m.jpg** with the results in step 2.
 
-## 3. Image PCA  Analysis
+## 2. Image PCA  Analysis
 The dataset you need is under `Face_dataset/` directory, which contains 56×46 pixel face images of 40 different subjects (classes), and 10 images available for each subject. Note that, i_j.png means personi_imagej . Now you have to split the dataset into two subsets (i.e., training and test sets). The first subset contains the **first 6 images** of each subject, while the second subset include the remaining images. Thus, a total of 6 × 40 = 240 images are in the training set, while 160 images in the test set.
 
 1. Perform PCA on the **training set**. Save the **mean face** and the first **three** eigenfaces.

diff --git a/Lab_2/Readme.md b/Lab_2/Readme.md
@@ -18,7 +18,7 @@ This is part of HW3 in NTU Computer Vision: from Recognition to Geometry (CV) co
 You can only use the python packages below:
  - python standard library
  - numpy
- - cv2 (opencv)
+ - cv2 (opencv) excluding `cv2.findhomography`
 
 ## Unwarp the screen