Kategorien durchsuchen
Entdecken
Fiverr Pro
Deutsch
$
USD
I will do object detection, yolo, computer vision, opencv projects in python
Python AI and Machine Learning Developer for Projects and Automation
Hi, I’m Hamza Shakil from the Pakistan, specializing in AI automations, machine learning, object detection, image classification, chatbots, and Python-based AI/ML solutions. With part-time experience ...
Über diesen Service
Need a custom object detection or computer vision solution? I'll train it end to end for you.
Services I offer:
- Custom object detection using YOLOv8, YOLOv5, YOLO-NAS (SOTA models)
- Image classification, segmentation, pose estimation
- OpenCV pipelines for real-time video processing
- Data preprocessing, augmentation, and annotation
- Model training, fine-tuning, evaluation (mAP, precision, recall)
- FastAPI / Flask deployment with REST endpoints
- Docker containerization
- ONNX and TFLite export for edge deployment
Use cases I've worked with:
Defect and quality inspection (manufacturing), people and vehicle detection (surveillance, traffic), retail shelf analytics, crop and wildlife monitoring, drone footage analysis, medical imaging, custom document region detection, and more.
Tech stack: Python, PyTorch, TensorFlow, YOLOv8, OpenCV, FastAPI, Docker
I'm a final-year CS student. My Final Year Project is a real-time YOLOv8 detection system with a full React enterprise dashboard, so this is my technical home turf.
Scope: Basic trains on your labeled data (up to 500 images). Standard adds labeling (200 images) + API. Premium is full service. Beyond limits: $0.15/image
APIs:
Microsoft Computer Vision AI
Expertise:
Objekterkennung
•
Ausreißererkennung
•
Schätzung der Pose
Programmiersprache:
Python
Tools:
Jupyter-Notizbuch
•
opencv
•
tensorflow
•
SimpleCV
•
PyTorch
Frameworks:
scikit-learn
•
keras
•
PyTorch
•
Panda
Mein Portfolio
FAQ
Which YOLO version do you use for object detection?
YOLOv8 by default (fast, accurate, easy to deploy). YOLOv5 and YOLO-NAS available if your use case needs them.
Can you also handle image classification or segmentation?
Yes — image classification with CNN backbones (ResNet, EfficientNet) and instance/semantic segmentation with YOLOv8-seg or Mask R-CNN. Message me your use case.
What accuracy or mAP can I expect?
Depends on dataset size and quality. With 500+ clean labeled images, mAP@0.5 of 0.7-0.85 is typical. Small or occluded objects reduce this. Honest estimate before you order.
Do you handle data labeling and annotation?
Yes — Standard tier includes up to 200 labeled images, Premium up to 500. Larger jobs at $0.15/image as a custom add-on.
Can you deploy the model with FastAPI or Docker?
Yes — Standard includes FastAPI inference endpoint, Premium includes full Docker cloud deployment.
