search

BradLarson / GPUImage2

GPUImage 2 is a BSD-licensed Swift framework for GPU-accelerated video and image processing.
BSD 3-Clause "New" or "Revised" License
4.88k stars 611 forks source link

Hough Transform for Line Detection #50

Open maxcan opened 8 years ago

maxcan commented 8 years ago

It was in the first GPUImage but hasn't been ported over yet. This would also require the Parallel Coordinates transform be ported over.

maxcan commented 8 years ago

started work on this here: https://github.com/maxcan/GPUImage2/tree/mc01.1-add-hough-line-detection

very preliminary, but any thoughts / feedback / suggestions welcome.

davies-w commented 8 years ago

Im working on an archery scoring program, and I'd LOVE to see if I can test this. I'm a complete image processing and iOS/Mac newb, so any help you can give to me to get this up and running on a test image would be great.

natew commented 6 years ago

This would be awesome to see. We're working on a screen-based OCR system, this would be a key piece in detecting lines of text! (Combined with being able to attach GPUImage to a screen recording AVCaptureSession).

johndpope commented 6 years ago

distantly related - https://github.com/BradLarson/GPUImage/issues/2208

konstiantyn commented 5 years ago

I'm working on driver license detection and recognition for OCR system. I'm capturing card image using AVcapturesession and trying rectangle detection by using Hough transform. So I need preprocess captured image by using GPUImage2 library in Swift 5.0, but I can't build it in Swift 5.0. there are so many errors.

davies-w commented 5 years ago

Ase, not much help on using Hough for this, but there's probably other solutions as there are a lot of CC capture apps right? EG -- they're just scanning the barcode!

https://scannrapp.com

W

On Thu, Apr 11, 2019 at 7:48 AM Ase Naritoshi notifications@github.com wrote:

I'm working on driver license detection and recognition for OCR system. I'm capturing card image using AVcapturesession and trying rectangle detection by using Hough transform. So I need preprocess captured image by using GPUImage2 library in Swift 5.0, but I can't build it in Swift 5.0. there are so many errors.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/BradLarson/GPUImage2/issues/50#issuecomment-482145926, or mute the thread https://github.com/notifications/unsubscribe-auth/AGP1RnruFRnFZx2Mggsi3Eiykmtty787ks5vf0s0gaJpZM4JY9Jf .

davies-w commented 5 years ago

@Ase, I found this that might help you? https://github.com/garnele007/SwiftOCR no experience, but I'd have started with a higher level library. I'm lurking here as I'm hoping one day to detect arrow shafts in a target, which is pretty niche, but it feels like card scanning is a "solved" problem at this point, so unless you're doing it for a learning project, I'd just find something that's targeted at your problem.

On Thu, Apr 11, 2019 at 7:48 AM Ase Naritoshi notifications@github.com wrote:

I'm working on driver license detection and recognition for OCR system. I'm capturing card image using AVcapturesession and trying rectangle detection by using Hough transform. So I need preprocess captured image by using GPUImage2 library in Swift 5.0, but I can't build it in Swift 5.0. there are so many errors.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/BradLarson/GPUImage2/issues/50#issuecomment-482145926, or mute the thread https://github.com/notifications/unsubscribe-auth/AGP1RnruFRnFZx2Mggsi3Eiykmtty787ks5vf0s0gaJpZM4JY9Jf .