Real-time Throw Detection Model Developer

I need a freelancer with expertise in Computer Vision, Deep Learning, and Video Analysis to build a model that can detect whether an individual’s throw (e.g., waste or object) goes inside or outside a dustbin from CCTV or camera footage.

Key Requirements:

- Develop a throw detection model using deep learning (preferred frameworks: TensorFlow, PyTorch, or similar).
- Ability to detect and classify a throw action in real-time video.
- Identify and label whether the throw lands inside the dustbin or outside it.
- Optimize for accuracy and efficiency (should work without high computational cost).
- Deliver clean, well-documented code and workflow for training and inference.
- Provide support for integrating the model with a cloud-based system later (optional but preferred).

Skills Required:

- Computer Vision (OpenCV, Mediapipe, or equivalent).
- Action Recognition Models (I3D, SlowFast, or custom CNN/LSTM/Transformer).
- Deep Learning frameworks (TensorFlow, PyTorch).
- Experience with object tracking and pose estimation is a plus.

Deliverables:

- Trained throw detection model.
- Source code with documentation.
- Sample demo on test video(s).