CNN Camera Pose Estimation (install remotely on our computer)
Budget: $750 – $1,500 NZD
I need an end-to-end CNN pipeline developed in PyTorch. The goal is to estimate camera pose from head images using synthetic data.
Requirements:
- Input: 40,000 synthetic head images in PNG format with camera coordinates.
- Output: Camera pose estimation in JSON format.
- Script: Validate by overlaying camera angle on test photo (RGB), showing origin x,y,z of the head.
Ideal Skills and Experience:
- Proficiency in PyTorch
- Experience with CNNs and image processing
- Strong background in computer vision
- Ability to write validation scripts
- Familiarity with JSON output format
Please ensure the solution is installable and runnable on our workstation for future training with new data.
Requirements:
- Input: 40,000 synthetic head images in PNG format with camera coordinates.
- Output: Camera pose estimation in JSON format.
- Script: Validate by overlaying camera angle on test photo (RGB), showing origin x,y,z of the head.
Ideal Skills and Experience:
- Proficiency in PyTorch
- Experience with CNNs and image processing
- Strong background in computer vision
- Ability to write validation scripts
- Familiarity with JSON output format
Please ensure the solution is installable and runnable on our workstation for future training with new data.