Comment by Social Finesse on "Image & Video Background Removal using Deep Learning"

Hey man I need your help! Great work on this by the way it's amazing for us video editors, but i'm having trouble near the last part of the code. I'm definitely not an avid coder here at all so let me know what I'm doing wrong. Everything looks like it activated correctly but when I get towards this part:

Removing the background from all frames and saving them

When I try to run that code, this error comes up:

NameError Traceback (most recent call last)

<ipython-input-7-ad0006e31003> in <module>() 32 ht,wd,l=final_img.shape 33 sz=(wd,ht) ---> 34 img_array.append(final_img)

NameError: name 'img_array' is not defined

From that point on, I am totally stuck here and can't finish the process because I don't know what this means or how to fix it. However, from what I can see, everything before that has activated successfully.

If you know how to fix this part for me, I'd greatly appreciate it! Thank you

I met the same problem:"NameError: name 'img_array' is not defined".①At first,I thought the package PIL was not imported successfully.I added "from PIL import Image" to the first line of the cell.However,it didn't work.②Then,given that the problem concerning "img_array",I added "img_array = []" before line 34(around line 29).It seemed to start working without error,the size of output.mp4 is only about 20K.After a while,I found that I fogot the "append()" function is used to add a new object to the end of the list.The addition of "img_array = []" was wrong.It is wise to add it just before the "for" loop(around line 2).③ However,new error jumped out："error: (-209:Sizes of input arguments do not match) The operation is neither ‘array op array’ (where arrays have the same size and the same number of channels), nor ‘array op scalar’, nor ‘scalar op array’ in function ‘cv::arithm_op’".It means that the sizes do not match?After I deleted "img_array = []" and "from PIL import Image",the cell began to work properly. It is a strange error that does not report an error when used alone.I don't know exactly what went wrong, so I wrote down the process of solving it. I hope it will help you. Good luck!

Here is my cell: for i in range(count):

#u2netreusult u2netresult=cv2.imread('/content/gdrive/MyDrive/background_removal_DL/test_data/videos/u2net_results/input'+str(i)+'.png')

#original original=cv2.imread('/content/gdrive/MyDrive/background_removal_DL/test_data/videos/input_frames/input'+str(i)+'.png')

#subimage subimage=cv2.subtract(u2netresult,original) cv2.imwrite('/content/gdrive/MyDrive/background_removal_DL/test_data/videos/output_frames/output'+str(i)+'.png',subimage)

#subimage subimage=Image.open('/content/gdrive/MyDrive/background_removal_DL/test_data/videos/output_frames/output'+str(i)+'.png')

#original original=Image.open('/content/gdrive/MyDrive/background_removal_DL/test_data/videos/input_frames/input'+str(i)+'.png')

subimage=subimage.convert("RGBA") original=original.convert("RGBA")

subdata=subimage.getdata() ogdata=original.getdata()

newdata=[] for i in range(subdata.size[0]*subdata.size[1]): if subdata[i][0]==0 and subdata[i][1]==0 and subdata[i][2]==0: newdata.append((255,255,255,0)) else: newdata.append(ogdata[i]) subimage.putdata(newdata) subimage.save('/content/gdrive/MyDrive/background_removal_DL/test_data/videos/output_frames/output'+str(i)+'.png',"PNG")

final_img=cv2.imread('/content/gdrive/MyDrive/background_removal_DL/test_data/videos/output_frames/output'+str(i)+'.png') ht,wd,l=final_img.shape sz=(wd,ht) img_array.append(final_img)

Search Hashnode