00:42:21	Andrew.French:	---------------------------------------------------------------------------
NameError                                 Traceback (most recent call last)
<ipython-input-4-8e6c68c81e44> in <module>
----> 1 I_pepper_gray_resize = skimage.transform.resize(I_pepper_gray,(128,128))
      2 plt.figure(figsize=(25,25))
      3 plt.imshow(I_pepper_gray_resize,cmap='gray')
      4 plt.axis('off')
      5 plt.title('Peppers, gray, resized')

NameError: name 'I_pepper_gray' is not defined

00:43:43	Andrew.French:	ok had to restart the script
00:43:45	Laura Boucheron:	I_pepper = np.asarray(imageio.imread('peppers.png'))
I_pepper_gray = skimage.color.rgb2gray(I_pepper)
00:44:05	Andrew.French:	thx
00:44:34	Kerrie Geil:	may want to do Kernel > Restart Kernel, then put your cursor in the block you left off on and then do Run > Run All above Selected Cell
00:46:47	Jennifer Woodward-Greene:	origH, origW = I_pepper_gray.shape
print(origH, origW)
aspect = origW/(origH*1.0)
print(aspect)
resizedH = int(25/aspect)
resizedW = int(resizedH*aspect)
print(resizedH, resizedW)

I_pepper_gray_myresize = skimage.transform.resize(I_pepper_gray,(resizedH, resizedW))

#resized dimensions show
plt.figure(figsize=(5,5))
plt.axis('off')
plt.imshow(I_pepper_gray_myresize,cmap='gray')
00:46:57	Jennifer Woodward-Greene:	oops. i did not crop...
00:47:07	Maria Laura Cangiano:	I_pepper_gray_crop = I_pepper_gray[:,64:448]
I_pepper_gray_resize = skimage.transform.resize(I_pepper_gray_crop,(256,256))
00:48:26	Maximilian Feldman:	I would pad the image with zeros to make it square (500X500). Then I would resize the image. As Jennifer did.
00:48:49	Maximilian Feldman:	I’m still trying to figure out how to do that.
00:49:01	Maximilian Feldman:	Maybe numpy.pad()
00:49:38	YAKOV PACHEPSKY:	Cut the 394x384 and then resized to 256x256
4
00:51:22	zhanyou xu:	is there parameter like keep_ratio= "Yes"?
00:52:37	Jerry M:	pick a corner and make a square box, then resize
00:52:52	Jerry M:	import matplotlib.patches as patches
box_corner = (75,120)
box_length = 100
cropped_image = I_pepper_gray[box_corner[1]:box_corner[1] + box_length, box_corner[0]:box_corner[0] + box_length]
I_pepper_gray_resize = skimage.transform.resize(cropped_image,(256,256))
plt.figure(figsize=(20,20))
fig, (ax1, ax2) = plt.subplots(1, 2)
rect = patches.Rectangle(box_corner ,box_length,box_length,linewidth=1,edgecolor='w',facecolor='none')
ax1.add_patch(rect)
ax1.imshow(I_pepper)
ax2.imshow(I_pepper_gray_resize,cmap='gray') 
ax1.axis('off') 
ax2.axis('off') 
plt.show()
00:59:43	zhanyou xu:	I_camera_rgb = skimage.color.gray2rgb(I_camera)

I_camera_gray_resize2 = skimage.transform.resize(I_camera_rgb,(384,512, 3))
plt.figure(figsize=(5,5)) 
plt.imshow(I_camera_gray_resize2,cmap='gray') 
plt.axis('on') 
plt.title('Peppers, gray, resized')
plt.show()

00:59:51	zhanyou xu:	Is this what you want us to do? 
01:03:33	Kale Harbick:	I_pepper_gray_crop = I_pepper_gray[128:384,64:320]
01:03:50	Kale Harbick:	oops
I_camera_resize = skimage.transform.resize(I_camera,(512,512))
I_camera_resize = I_camera_resize[64:448,0:512]
01:03:53	Jennifer Woodward-Greene:	happily working
01:04:40	Lucas Heintzman:	My "%who" command is no longer returning metadata information, all I get is a list of object names. Have I exceeded some limit?
01:05:04	Curtis Ransom:	should be %whos
01:05:41	Lucas Heintzman:	Corrected, ty.
01:05:52	Jennifer Woodward-Greene:	#_camera = I_camera[0:256,0:256]
plt.figure()
plt.axis('off')
plt.imshow(I_camera, cmap='gray')

pepperaspect = 384/512
print('pepper', pepperaspect)

I_camera_cropped = I_camera[0:256,0:192]

origH, origW = I_camera_cropped.shape
print(origH, origW)
aspect = origW/(origH*1.0)
print(aspect)
resizedH = int(512/aspect)
resizedW = int(resizedH*aspect)
print(resizedH, resizedW)
print('aspect new', resizedW/(resizedH*1.0))

I_camera_myresize = skimage.transform.resize(I_camera_cropped,(resizedH, resizedW))

#resized dimensions show
plt.figure()
plt.axis('off')
plt.imshow(I_camera_myresize,cmap='gray')
01:06:12	Andrew.French:	btw, workbook is really nice for teaching, but how practical in production?
01:06:24	Jerry M:	I_camera_resize = skimage.transform.resize(I_camera_rgb,(384,384))
full_size_image = np.zeros([384,512,3])
for i in range(I_camera_resize.shape[0]):
    for j in range(I_camera_resize.shape[1]):
        for k in range(I_camera_resize.shape[2]):
            if (j<384):
                full_size_image[i,j,k] = I_camera_resize[i,j,k]
            
plt.figure(figsize=(20,20)) 
plt.imshow(I_camera_resize,cmap='gray') 
plt.axis('off') 
plt.title('Peppers, gray, resized')
plt.show()
full_size_image.shape
01:08:56	Jerry M:	error in my code:
01:09:03	Jerry M:	

I_camera_resize = skimage.transform.resize(I_camera_rgb,(384,384))
full_size_image = np.zeros([384,512,3])
for i in range(I_camera_resize.shape[0]):
    for j in range(I_camera_resize.shape[1]):
        for k in range(I_camera_resize.shape[2]):
            if (j<384):
                full_size_image[i,j,k] = I_camera_resize[i,j,k]
            
plt.figure(figsize=(20,20)) 
plt.imshow(full_size_image,cmap='gray') 
plt.axis('off') 
plt.title('Peppers, gray, resized')
plt.show()
full_size_image.shape
01:11:48	Jennifer Woodward-Greene:	pepperaspect = 384/512
print('pepper', pepperaspect)

I_camera_cropped = I_camera[0:256,0:int(256*pepperaspect)]
01:12:14	YAKOV PACHEPSKY:	Stretch camera to 512x512, cut 384X512
01:16:10	Maria Laura Cangiano:	I_camera_resized = I_camera_rgb[32:224,:,:]
I_camera_resized = skimage.transform.resize(I_camera_resized,(384,512,3))
01:16:39	Jennifer Woodward-Greene:	first one
01:16:59	Lucas Heintzman:	First one: due to higher complexity of subject matter.
01:17:06	YAKOV PACHEPSKY:	Nodiffrence, really
01:37:21	Andrew.French:	how are the edges of the image handled, ie in the taper zone
01:37:25	Andrew Russ:	Blurs image, more affect on areas with fine detail and edges
01:37:30	Jerry M:	2 box blurs with differing cutoff frequencies

01:37:36	Jennifer Woodward-Greene:	Edges are less defined.
01:37:47	YAKOV PACHEPSKY:	why would you want to lose information?
01:40:19	Timothy:	could this be used when there are missing pixels or so?
01:40:24	Andrew.French:	yes
01:41:58	Jerry M:	Is texture always high frequency?
01:45:04	Kerrie Geil:	For folks that are new to Jupyter, you can also see help on different functions by showing the "contextual help" pane in your notebook. Type Ctl+I ("eye") to bring up the contextual help. Drag the contextual help tab around Jupyter Lab to place, for example, at the bottom of your screen so you can see your notebook and the contextual help. Then, when you place your cursor after a function you should be able to see the help in the contextual area

01:53:32	Jennifer Woodward-Greene:	h3 appears 3D, it has 'depth' look
01:53:38	Andrew Russ:	seems like they miss edges on the right
01:53:47	Jennifer Woodward-Greene:	yes
01:54:01	Maximilian Feldman:	h3 is vertical h4 is horizontal?
01:54:43	Maximilian Feldman:	h3 is finding difference along y-axis
01:56:35	Jennifer Woodward-Greene:	Because the distance between the coat and the stand is too close for 9x9?
02:05:52	Maria Laura Cangiano:	How do you know which kernel to use? Meaning which numbers do you pick for your filters?
02:07:29	Timothy:	I have heard of filtering using mean, median etc values. How do you specify those?
02:07:38	Elizabeth Chin:	is there any filter to enhance texture? for example, the edges of the grass would be emphasized in the camera image.
02:08:48	Maria Laura Cangiano:	yes, thanks
02:10:11	Timothy:	Sure thanks
02:12:25	Elizabeth Chin:	ok thanks!
02:12:54	Jennifer Woodward-Greene:	will you send your copy of the notebook today?
02:13:49	Kerrie Geil:	https://kerriegeil.github.io/NMSU-USDA-ARS-AI-Workshops/learn/
02:14:39	Jennifer Woodward-Greene:	Thanks!
02:17:21	Kerrie Geil:	--reservation=kerrie.geil_86 wednesday
02:18:18	Maria Laura Cangiano:	In the instructions in the tutorial the  -xvfz to extract the files from the tar files  has an extra z it's actually -xvf, or at least it was like that for me
02:19:17	YAKOV PACHEPSKY:	I concur with Maria
02:20:51	Maria Laura Cangiano:	I already downloaded the files. Thank you for the workshop today, it was very informative and perfectly organized. See you on Wednesday!
02:25:57	Elizabeth Chin:	i was able to untar and uncompress using: tar -xvzf filename (note the z and f are switched from the workshop website)
02:26:05	Jennifer Woodward-Greene:	I extracted Annotations no problem. 101_ObjectCategories.tar went to 101_ObjectCategories.tar.gz. Am I done?
02:26:21	Kerrie Geil:	You are done Jennifer
02:26:42	Jennifer Woodward-Greene:	Thanks... no folder for the second one (like appears in your directory)
02:27:36	Jennifer Woodward-Greene:	I am not able to open the tar.gz
02:29:08	Mohamad Moussa:	what is supposed to be in the 101_objectCategoreis folder?
02:30:30	Jennifer Woodward-Greene:	I do not show a second folder... it looks different than Suzy's!
02:32:15	Jennifer Woodward-Greene:	can you put the command here? My powershell is disabled for scripts...
02:40:13	Suzy Stillman:	tar -xvf 101_ObjectCategories.tar
02:47:42	Jennifer Woodward-Greene:	Thanks all of you!!! See you Wednesday.