Sfoglia categorie
Esplora
Fiverr Pro
Italiano
$
USD
Sembra che questo servizio sia in sospeso
I will computer vision ai mobile app for sports player detect bytettrack deepsort yolo
Regno Unito
Parlo Inglese, Spagnolo
Full Stack AI ML NLP Engineer Data Scientist Computer Vision
Full-Stack Developer & AI Engineer with 11+ years experience. Python, React, Django, Java, C++, C#, Spring Boot, AWS/Azure/GCP expert. Int...
Informazioni su questo servizio
Want a mobile app that detects and tracks players in sports footage with production-grade accuracy? I develop computer vision mobile apps that combine YOLO detection with ByteTrack / DeepSORT multi-object tracking to deliver stable player IDs, trajectory analysis, and exportable analytics. The app supports live camera or video input, overlays bounding boxes and player IDs, computes speed/trajectory metrics, and outputs structured data (CSV/JSON or live API). Ideal for coaches, analysts, broadcasters, and sports tech teams working on football, basketball, tennis, athletics, climbing, or combat sports.
Services We Offer (quick list)
- YOLO model integration for player & object detection
- ByteTrack or DeepSORT for multi-player consistent ID tracking
- Pose estimation integration (MediaPipe/OpenPose)
- Mobile optimization (TFLite / Core ML conversion & quantization)
- Live camera overlay: bounding boxes, IDs, confidence, metrics
- Player trajectory, speed, and heatmap generation
- Export: CSV / JSON logs, annotated video, real-time API/WebSocket
- Multi-camera or streamed input (optional)
- Backend streaming endpoint (FastAPI) and dashboard hooks
- Dataset labeling & custom model fine-tuning
Contact Us
Linguaggio di programmazione:
Python
•
R
•
Colab
•
NoSQL
•
MLflow
Framework:
Scikit-learn
•
DeepPy
•
Google ML Kit
•
PyTorch
•
Panda
FAQ
App UI and frontend
Flutter / React Native / Swift / Kotlin
Do you need my footage to start?
Yes, sample videos (varied angles, lighting) greatly improve accuracy and help me tune models.
Will this run in real time on mobile?
Yes, with optimized models (TFLite/Core ML) and appropriate device hardware. Server inference is an option for heavier models.
Can you distinguish players reliably?
ByteTrack/DeepSORT provide consistent ID tracking; accuracy increases with better camera views and dataset fine-tuning.
Can you detect events (passes, shots, tackles)?
Yes, event detection is available as an add-on using pose/motion features and ML classifiers trained on labeled events.
Do you provide dataset labeling?
Yes, dataset annotation and custom training are available (quoted separately).
Can you integrate with my dashboard or analytics pipeline?
Yes, I provide CSV/JSON exports, REST API/WebSocket streaming, and dashboard integration guidance.
