We are seeking a Computer Vision Specialist to deliver a specific, well-defined computer vision task, not a long-term team role . The ideal candidate will have experience in developing computer vision algorithms and solutions, with a focus on image processing and object recognition. You will be responsible for implementing and optimizing computer vision models, and collaborating with our engineering team to integrate these solutions into our product. If you have a passion for technology and innovation, we would love to hear from you!
Scope of Work:
1- Analyze production-line images
2- Detect all visible cup bases in each image
3- Accurately count cup bases per image
4- Handle real-world challenges such as:
- Lighting variations
- Noise and blur
- Reflections from plastic surfaces
- Minor perspective changes
5- Deliver a fully automated and offline solution
No specific framework or algorithm is mandated.
Input:
1- RGB images from an industrial production line
2- Images may include edge cases and challenging conditions
Outputs:
1. Visual Output
- Annotated images showing detected cup bases
- Clear total count displayed per image
2. Data Output
• CSV / JSON (or equivalent) containing:
- Image ID
- Total detected cup count
- Optional confidence or validation metrics
Critical Accuracy Requirement:
✅ Minimum Accuracy: 99.5%
Accuracy calculation:
Accuracy = (Correct Detections / Ground Truth Count) × 100
Acceptance threshold:
Maximum 1 incorrect detection per 200 detections
Any result below 99.5% accuracy will be rejected
Validation:
Tested on previously unseen images
Includes normal cases and edge cases
Results must be:
Fully automated
Repeatable
Free from manual correction
Constraints:
No manual interaction during runtime
Must work offline
No external services or internet usage
Deliverables:
Source code and Executable system (ready to run)
Usage instructions (execution only – no algorithm disclosure required)
Sample output images
Formal accuracy report
Ideal Freelancer:
Strong experience in Computer Vision
Proven background in industrial inspection systems
Excellent handling of real-world noisy data
Ability to deliver production-level accuracy
Apply Now
Apply Now