Development of Document Contour Recognition Algorithm
Budget: $250 – $750 USD
Okay, here is a draft job posting in English for the freelancer role you described. You can adapt this for specific platforms (like Upwork, Freelancer.com, LinkedIn, etc.).
Job Title: Freelance Algorithm/AI Developer for High-Precision, Real-time Document Boundary Detection on Edge Devices
Project Overview:
We are seeking a highly skilled and experienced freelance developer specializing in computer vision and potentially AI/Machine Learning. The goal of this project is to create a robust algorithm capable of accurately and rapidly detecting the precise boundaries (lines and corners/vertices) of documents within images captured by mobile devices. The solution must perform in real-time on edge devices (smartphones) and achieve high accuracy comparable to leading document scanning applications like CamScanner.
Responsibilities:
Design, develop, and implement an efficient algorithm (using traditional computer vision techniques, AI/ML models, or a hybrid approach) to precisely identify the corners and edges of various documents in diverse image conditions (lighting, angles, backgrounds, document types).
If an AI model approach is chosen, potentially handle data sourcing/augmentation, model training, evaluation, and optimization.
Optimize the algorithm/model for real-time inference speed and low resource consumption on typical mobile/edge hardware.
Ensure the output provides highly accurate coordinates for the document's corners/polygon.
Provide clean, well-documented, and potentially portable code (e.g., Python with OpenCV, C++, or a format suitable for mobile integration like TensorFlow Lite, ONNX, Core ML).
Collaborate via clear communication regarding progress, challenges, and potential solutions.
Required Skills and Experience:
Strong background in Computer Vision and Image Processing.
Proven experience in developing algorithms for object detection, segmentation, or specifically document detection/scanning.
Proficiency in relevant programming languages and libraries (e.g., Python, C++, OpenCV).
Experience with algorithm optimization for performance-critical applications, especially on mobile/edge devices.
Understanding of geometric transformations (e.g., perspective correction).
Ability to work independently and deliver high-quality results.
Good communication skills in English.
Highly Desirable (Preferred Qualifications):
Experience with developing and deploying Machine Learning models (especially CNNs) for mobile/edge devices (TensorFlow Lite, PyTorch Mobile, Core ML, ONNX Runtime, etc.).
Proven ability and willingness to demonstrate a proof-of-concept or a working demo of your proposed approach before project engagement. This is a strong preference.
Experience or capability in reverse engineering existing algorithms or applications is a significant plus.
Deliverables:
Optimized source code for the final algorithm/model.
Clear documentation explaining the approach, usage, and integration steps.
(If applicable) Trained model files and details on the training process/data.
Job Title: Freelance Algorithm/AI Developer for High-Precision, Real-time Document Boundary Detection on Edge Devices
Project Overview:
We are seeking a highly skilled and experienced freelance developer specializing in computer vision and potentially AI/Machine Learning. The goal of this project is to create a robust algorithm capable of accurately and rapidly detecting the precise boundaries (lines and corners/vertices) of documents within images captured by mobile devices. The solution must perform in real-time on edge devices (smartphones) and achieve high accuracy comparable to leading document scanning applications like CamScanner.
Responsibilities:
Design, develop, and implement an efficient algorithm (using traditional computer vision techniques, AI/ML models, or a hybrid approach) to precisely identify the corners and edges of various documents in diverse image conditions (lighting, angles, backgrounds, document types).
If an AI model approach is chosen, potentially handle data sourcing/augmentation, model training, evaluation, and optimization.
Optimize the algorithm/model for real-time inference speed and low resource consumption on typical mobile/edge hardware.
Ensure the output provides highly accurate coordinates for the document's corners/polygon.
Provide clean, well-documented, and potentially portable code (e.g., Python with OpenCV, C++, or a format suitable for mobile integration like TensorFlow Lite, ONNX, Core ML).
Collaborate via clear communication regarding progress, challenges, and potential solutions.
Required Skills and Experience:
Strong background in Computer Vision and Image Processing.
Proven experience in developing algorithms for object detection, segmentation, or specifically document detection/scanning.
Proficiency in relevant programming languages and libraries (e.g., Python, C++, OpenCV).
Experience with algorithm optimization for performance-critical applications, especially on mobile/edge devices.
Understanding of geometric transformations (e.g., perspective correction).
Ability to work independently and deliver high-quality results.
Good communication skills in English.
Highly Desirable (Preferred Qualifications):
Experience with developing and deploying Machine Learning models (especially CNNs) for mobile/edge devices (TensorFlow Lite, PyTorch Mobile, Core ML, ONNX Runtime, etc.).
Proven ability and willingness to demonstrate a proof-of-concept or a working demo of your proposed approach before project engagement. This is a strong preference.
Experience or capability in reverse engineering existing algorithms or applications is a significant plus.
Deliverables:
Optimized source code for the final algorithm/model.
Clear documentation explaining the approach, usage, and integration steps.
(If applicable) Trained model files and details on the training process/data.