willbrennan / skindetector Goto Github PK

could you guide me on how I can do this

I'm using this for video redaction of body parts. Could you please add a way to increase sensitivity?

Hello William! Congratulations, this is a wonderfull script. Well I found something interesting and related with the issue #11 . Every time the program runs, it outputs a mask with only one chanel. That's why it generates black and white images. This is intentional? Thank you very much for this library!

I am runing opencv 3.0, python 3.5 in ubuntu 16.04

When I run:
python main.py ~/MovieProcessing/alvaro/delete/p6-1.png --display

here is the image

I get the error

OpenCV Error: Sizes of input arguments do not match (The operation is neither 'array op array' (where arrays have the same size and type), nor 'array op scalar', nor 'scalar op array') in binary_op, file /home/aulloa/src/opencv/modules/core/src/arithm.cpp, line 225
Traceback (most recent call last):
File "main.py", line 197, in
img_msk = process(img_col, args=args)
File "main.py", line 188, in process
return detector.process(image)
File "main.py", line 125, in process
self.get_mask_rgb(img)
File "main.py", line 61, in get_mask_rgb
msk_rgb = cv2.bitwise_and(mask_a, mask_b)
cv2.error: /home/aulloa/src/opencv/modules/core/src/arithm.cpp:225: error: (-209) The operation is neither 'array op array' (where arrays have the same size and type), nor 'array op scalar', nor 'scalar op array' in function binary_op

The link in Based Upon section is no longer available, please check it.

Thank you so much for this repository, I'm curious about your skin detection library and having issues installing it. Please check out the screen shot attached.

I'd appreciate if you can please share me your email address for any potential collaboration. My email address is contact(at)phanig.com

Yo, Appreciate your work.
However , when I try to use , I cannot produce the image shown in the demonstration.
My code :
image = cv2.imread("path_of_img")
mask = skin_detector.process(image)
cv2.imwrite(os.path.expanduser("path_destination_result"), mask)

The Image Input:

The Image Output:

However, I want to have the result output image to keep its original skin color , Please Help , Thanks so much !

I am using opencv 4.2 and not sure whether it's a bug on early version.

I found that the returned HSV mask was almost empty (all zero), so I went deep into the codes and found 2 problems.

1. The default color space of opencv is BGR, so the second parameter of cvtColor should be COLOR_BGR2HSV/COLOR_BGR2YCrCb. Accordingly, the threshold of YCrCb should also change to [90, 130, 100] ~ [230, 180, 120] (swap Cr and Cb channel)

2. The upper threshold of HSV may be [20, 150, 255]. H channel from 0 to 120 is too wide. The range of H channel in opencv is 0~180.

Is there any way to produce the RGB values of the skin in the console output? Something similar to how the output is produced as segments here But, I'm only wanting to produce a single RGB value for the skin. Any help would be highly appreciated!