AI-Based Line Detection, Linking, and Rectangle Formation
Budget: $250 – $750 CAD
I'm looking for an AI solution that can effectively process scanned documents. This solution will detect segmented lines within these images, link and straighten these lines while maintaining the original structure, and finally, form rectangles from these linked lines. It's important that the solution has the capability to handle multiple lines on a single rectangle side.
Key aspects of the project include:
- Input: The primary input for this project will be images containing segmented lines, generally scanned documents.
- Output: The desired output will be the detected and linked rectangles in vector format.
- AI Approach: I prefer a deep learning approach, ideally utilizing Convolutional Neural Networks (CNNs) or Transformers, combined with traditional methods from OpenCV.
- Performance: The solution must be robust enough to handle noise, distortions, and varying line thicknesses.
Ideal skills and experience for this project:
- Proficiency in Python, particularly with TensorFlow as the preferred framework.
- Demonstrable experience in AI-driven image processing, specifically with deep learning and traditional computer vision techniques.
- Ability to train on custom datasets and refine models over time will be a bonus.
Please apply if you have relevant expertise and include details of your approach, previous projects, and experience in the field.
Key aspects of the project include:
- Input: The primary input for this project will be images containing segmented lines, generally scanned documents.
- Output: The desired output will be the detected and linked rectangles in vector format.
- AI Approach: I prefer a deep learning approach, ideally utilizing Convolutional Neural Networks (CNNs) or Transformers, combined with traditional methods from OpenCV.
- Performance: The solution must be robust enough to handle noise, distortions, and varying line thicknesses.
Ideal skills and experience for this project:
- Proficiency in Python, particularly with TensorFlow as the preferred framework.
- Demonstrable experience in AI-driven image processing, specifically with deep learning and traditional computer vision techniques.
- Ability to train on custom datasets and refine models over time will be a bonus.
Please apply if you have relevant expertise and include details of your approach, previous projects, and experience in the field.