You will get a traffic analysis application
Top Rated
Top Rated
Project details
You will get a Real-time Traffic Violation Detection System, an innovative end-to-end solution designed to detect traffic violations accurately and efficiently in video footage.
Features:
• Vehicle Speed and Distance estimation
• Vehicle Tracking
- Tracks vehicles even when plates are not readable in subsequent frames.
• Plate Recognition:
- Detects plates using a YOLOv8 model.
- Reads plate numbers with EasyOCR.
- Links plate numbers with tracked vehicle IDs.
• Violation Detection:
- Detects predefined violations using speed and distance estimates.
- Stores details of violations, including costs and frames, in a CSV file.
• Violation Reporting:
- Generates video clips of violations for review.
- Provides detailed reports with all relevant information.
• Tools Used:
- Vehicle Speed and Distance Detection: Optical flow techniques.
- Vehicle Tracking: Deep SORT algorithm with PyTorch, supporting YOLO series models.
- Plate Recognition: YOLOv8 model and EasyOCR library.
- Violation Detection and Reporting: Custom scripts to detect violations and generate reports.
Features:
• Vehicle Speed and Distance estimation
• Vehicle Tracking
- Tracks vehicles even when plates are not readable in subsequent frames.
• Plate Recognition:
- Detects plates using a YOLOv8 model.
- Reads plate numbers with EasyOCR.
- Links plate numbers with tracked vehicle IDs.
• Violation Detection:
- Detects predefined violations using speed and distance estimates.
- Stores details of violations, including costs and frames, in a CSV file.
• Violation Reporting:
- Generates video clips of violations for review.
- Provides detailed reports with all relevant information.
• Tools Used:
- Vehicle Speed and Distance Detection: Optical flow techniques.
- Vehicle Tracking: Deep SORT algorithm with PyTorch, supporting YOLO series models.
- Plate Recognition: YOLOv8 model and EasyOCR library.
- Violation Detection and Reporting: Custom scripts to detect violations and generate reports.
Machine Learning Tools
ChatGPT, Deeplearning4j, Google AutoML, GPT-3, MLflow, NumPy, OpenCV, pandas, Python, Python Scikit-Learn, PyTorch, scikit-learn, SQLWhat's included
| Service Tiers |
Starter
$250
|
Standard
$500
|
Advanced
$1,500
|
|---|---|---|---|
| Delivery Time | 1 day | 2 days | 10 days |
Number of Revisions | 1 | 2 | Unlimited |
Number of Model Variations | 1 | 1 | 5 |
Number of Scenarios | 1 | 2 | 5 |
Number of Graphs/Charts | 10 | 20 | 50 |
Model Validation/Testing | |||
Model Documentation | - | ||
Data Source Connectivity | - | - | |
Source Code |
12 reviews
(12)
(0)
(0)
(0)
(0)
This project doesn't have any reviews.
MO
Michael O.
Jan 22, 2025
60 minute consultation
I had a great consultation with Mohamed, and I’m thoroughly impressed. He took the time to carefully explain everything, ensuring I understood each concept. It’s clear that he really knows his stuff and has a deep understanding of AI and computer vision. Highly recommend his expertise!
FA
Fahad A.
Jan 5, 2025
This contract will be for our AI solution platform and its interface for end users like doctors,
calm and focused communicator with punctual behavior,
deep technical knowledge
deep technical knowledge
AA
Abdullah A.
Jul 28, 2024
R&M AI
He was helpful and completed the necessary work, and even more.
JF
Jerry F.
Jul 26, 2024
Optimization Pt1
FA
Fernando A.
Apr 15, 2024
60 minute consultation
About Mohamed
Data Scientist | Computer Vision Engineer | LLMs developer
100%
Job Success
Asyut, Egypt - 3:38 am local time
I help startups and companies turn ideas into working AI products: from data preparation and model development to deployment and integration into real applications. I’ve worked as the sole ML engineer for a healthcare AI startup, built production-ready computer vision systems, and developed end-to-end ML pipelines for both research and industry use.
What I can help you with
🔹 Computer Vision: image classification, object detection, segmentation, tracking, video understanding, pose-based analysis, medical imaging AI, error analysis, model debugging, and data augmentation workflows.
🔹 Machine Learning Engineering: PyTorch model development, multi-GPU training, experiment tracking, reproducible workflows, model evaluation, failure analysis, performance optimization, ONNX, TensorRT, and Triton deployment.
🔹 AI Product Development: building MVPs powered by ML, FastAPI backends for model serving, integrating models into web or mobile products, Dockerized deployment, and cloud integration.
Selected experience
🚀 Built and deployed medical imaging models for a healthcare AI startup, including a chest X-ray classifier with 95.5% F1 and a dual-view breast cancer model with 85% F1 plus tumor localization.
🚀 Improved a satellite object detection system from 84.5% to 92.0% mAP@50 by replacing a YOLOv8 baseline with RF-DETR and building analysis and deployment workflows around it.
🚀 Developed a video computer vision pipeline for golf swing analysis using LSTM, MediaPipe, YOLOv8, OpenCV, FastAPI, and GCP.
🚀 Built ML workflows using Airflow, MLflow, ONNX, TensorRT, and Triton.
🚀 Contributed to the Ivy open-source library.
Tech stack
💻 Languages: Python, C++, SQL
🧠 Frameworks/Libraries: PyTorch, OpenCV, Transformers, TensorFlow, MediaPipe, FastAPI
⚙️ Deployment/Infrastructure: Docker, Airflow, MLflow, ONNX, TensorRT, Triton, Linux, GCP, AWS
🗄️ Other tools: PostgreSQL, Git, LangChain
How I work
✅ I care about more than just training a model. I focus on building systems that are accurate, practical, maintainable, and ready for real users.
✅ Whether you need help with an ML prototype, a deployed computer vision system, or improving an existing model, I can help move your project from idea to working solution.
📩 Send me a message with your project details, and let’s discuss how I can help.
Steps for completing your project
After purchasing the project, send requirements so Mohamed can start the project.
Delivery time starts when Mohamed receives requirements from you.
Mohamed works on your project following the steps below.
Revisions may occur after the delivery date.
Learning about the problem
Researching the problem and the previous solutions.
Searching for the previous models and datasets
Most of the time, there are open-source models and datasets related to the problem we are trying to solve.