Python Edge-Line Cleaner Script

Pending

Prize:

$150 AUD

Entries Received:

28 days, 7 hours to award

I need a small Python utility that can scan any PNG I feed it, detect every black-colored pixel that is connected to the outer frame of the image, and then erase those pixels while leaving every interior mark—text, boxes, circles, fine lines—fully intact. The files come in many sizes and there is no alpha channel to worry about, so you are free to work in a standard 24-bit space. Feel free to rely on OpenCV, Pillow, NumPy, scikit-image or any similar library as long as installation stays simple (pip install …). Command-line execution is ideal: point the script at a single file or an entire folder, have it generate a cleaned version in a chosen output directory, and log how many border pixels were removed for quick verification. A clip example image is provided that shows the original image, the items to remove, and what the finished item should look like. When submitting, please show the output of all the clips, including the ones which are not in the example image. If you do not show all the clips then your entry will be rejected. Acceptance criteria • A .py file with clear, commented code • Works on an arbitrary-sized PNG and keeps all non-edge content untouched • Command-line arguments: input path (file or folder) and output path • Read-me or inline docstring explaining dependencies and usage If you have another smart approach to edge-connected component removal, tell me about it—speed and reliability matter more than any specific library choice.

Featured

Guaranteed

Skills Required

Computer Vision

Image Analysis

Image Processing

NumPy

OpenCV

Python

Accepted File Formats

gif, jpeg, jpg, png

Clarification Board
No spam, self-promotion or advertisement is permitted.

David Yoga P.

@dvyoga14

2 days ago

Hi Contest Holder, please read my message via chat. Thanks!

Muhammad Ibrahim M.

@Ibrahimtayyab989

4 days ago

Hi, this is IT from SOFINT. Before we finalize the OpenCV flood-fill masking logic, how should the script handle valid interior elements (like long lines or text boxes) that accidentally touch the very edge of the frame—should the entire connected shape be erased, or is there a maximum thickness limit for the border lines? If our technical approach makes sense to you, let’s connect in the inbox so you can review our processed clip samples; once you are fully satisfied with the results and our company, we can safely proceed to the awarding phase. — S O F I N T ∿

Mir Sajid K.

@sajidfuuast

4 days ago

check entry #56

Zafar M.

@dataexpert18

5 days ago

I am working on it.

Muhammad I.

@ikhan9985

5 days ago

You are missing something #30

Muhammad Ibrahim M.

@Ibrahimtayyab989

6 days ago

I read the requirement but I have a few major questions I’d like to ask you. Would it be okay if we continue this conversation in the inbox?

Sihina S.

@sihinas2

6 days ago

I have out of entries. I can not submit more entries.

Sihina S.

@sihinas2

6 days ago

Sir, please come to chat. I have all the output images as you need.

Avinash S.

@AvinaHSinha7

6 days ago

Hi Michael, Can we connect over chat ?

Avinash S.

@AvinaHSinha7

6 days ago

Hi, Please check #28. I have also attached a drive link where you can find the output of all the clips. I have the script ready. I will send all the output images and the script.