
Open
Posted
•
Ends in 4 days
Paid on delivery
I have a Python-based gaze-tracking pipeline that combines OpenCV for video I/O, Dlib for facial landmarks, and lightweight TensorFlow / Torch models for inference. The core flow works, but three pain points remain: overall gaze accuracy, occasional landmark drift afton screen, and sluggish frame rates when everything runs on a CPU-only machine. What I need now is a clean refactor or patch set that: • Tightens gaze estimation accuracy without sacrificing latency. • Eliminates the drifting landmarks issue so that the eye region stays correctly aligned over long clips. • Pushes real-time speed on an average CPU (no GPU assumption) by profiling bottlenecks and applying vectorised or threaded alternatives. Feel free to introduce different landmark detectors, filtering, model quantisation or any other optimisation that you can justify in code. I’ll test the result on my existing data set; if the gaze error drops and the FPS climbs, the job is done. Deliverables: • Updated Python scripts or modules with clear inline comments. • A short README noting dependency versions and the exact commands I should run. • Before/after metrics: accuracy (e.g., angular or pixel error) and FPS on CPU. If you can demonstrate tangible gains quickly, we can move on to additional refinements or feature ideas.
Project ID: 39739512
5 proposals
Open for bidding
Remote project
Active 2 days ago
Set your budget and timeframe
Get paid for your work
Outline your proposal
It's free to sign up and bid on jobs
5 freelancers are bidding on average ₹880 INR for this job

I am a seasoned software developer with 13 years of experience, holding a degree from IIT Delhi. My expertise aligns perfectly with the required skills for your project. I have successfully delivered complex solutions across diverse domains with a focus on quality and scalability. I bring strong problem-solving ability, hands-on technical depth, and client-centric delivery. I am confident I can add value to your project and deliver results within timelines.
₹1,000 INR in 7 days
0.4
0.4

?I TREAT EVERY PROJECT AS IF IT WERE MY OWN — WITH CARE, PRECISION AND ACCOUNTABILITY. Looking for speed, reliability, and quality? That’s exactly what I bring. I understand the importance of tightening gaze accuracy, eliminating landmark drift, and improving frame rates on CPU-only machines for your Python gaze-tracking project. I specialize in optimizing code for performance while maintaining accuracy, ensuring a seamless and efficient user experience. While I am relatively new to Freelancer, I have extensive experience with similar projects off-site. If you're looking for someone who treats your project like their own and delivers beyond expectations, I'd love to discuss further about your project! Regards, Praven
₹800 INR in 14 days
0.0
0.0

As a Full-Stack Developer with a special focus on Python, I am confident in my ability to deliver tangible improvements to your gaze-tracking pipeline. I understand the limitations you're facing regarding accuracy, landmark drift, and frame rate sluggishness; and I'm eager to bring my problem-solving skills to the table to address these pain points. Efficiency and precision are key in any project I take on, and that's exactly what you need here. Moreover, I'm no stranger to optimizing code for maximum performance, especially on CPU-only machines. From identifying bottlenecks and threading alternatives to model quantization or trying new landmark detectors, I'm ready to employ any optimization justified by code analysis to ensure tightened gaze estimation accuracy and improved frame rates without sacrificing latency. My approach has always been focused on delivering clean, maintainable, and scalable code - all ingredients necessary for successful project completion aligned with your goals. With a proven background in developing real-world applications with quantifiable improvements for businesses across various sectors, I'm confident I can do the same for your Python gaze tracking pipeline. Let's connect and discuss details on delivering a solution tailored uniquely for you!
₹800 INR in 4 days
0.0
0.0

Hello, With my experience in Python, OpenCV, and TensorFlow/Torch, I’m well-suited to optimize your gaze-tracking pipeline. I specialize in improving performance and accuracy, particularly in real-time applications, and have worked with both CPU and GPU environments to ensure smooth, high-speed inference. My focus will be on tightening gaze estimation, fixing landmark drift, and maximizing CPU performance without sacrificing accuracy. In a previous project, I worked on optimizing an object-tracking system that used Dlib for facial landmark detection, where I reduced processing time by over 30% through code refactoring and threading. I also implemented better landmark tracking using advanced filters and quantization, which significantly improved the tracking consistency and speed on CPU. These improvements resulted in smoother performance and more accurate results, much like the goals for your project. I’m excited about the opportunity to work on this, as I enjoy solving performance-related challenges in computer vision. I look forward to providing a clean solution with measurable improvements in both gaze accuracy and FPS, making your pipeline more efficient and reliable.
₹1,000 INR in 7 days
0.0
0.0

With my broad skill set and experience in Full-Stack Development and AI Integration, I'm keenly interested in optimizing your Python Gaze Tracking project. My successful development of 4+ live applications with measurable business impact speaks volumes about my commitment to delivering solutions that solve real problems - the same commitment I will bring to the table for your project. My skills in Python, AI/ML (including OpenCV and Computer Vision), as well as my experience in integrating AI into existing systems make me a strong candidate for ensuring not just optimized gaze tracking but also maintaining requisite latency. As an award-winning developer, I consistently buttress my code quality and this, coupled with providing you refactored Python scripts with clear inline comments, will make further modifications should you require them seamless. Moreover, I understand the criticality of speed on a CPU since you aren't making any GPU assumptions. Profiling bottlenecks and applying potent alternatives is something I'm well-versed with as I've recently built an AI test case generator with multi-framework support where optimization was paramount. We can discuss updating landmark detectors, filtering or even model quantisation if required to further improve your system. Let's optimize your gaze-tracking system while preserving real-time speed on your average CPU; together we can achieve this!
₹800 INR in 3 days
0.0
0.0

Greater Noida, India
Member since Feb 10, 2020
$50 USD
$50 USD
$30-250 USD
$10-30 USD
₹1500-12500 INR
₹75000-150000 INR
₹400-750 INR / hour
$30-250 USD
$10-30 USD
₹800-1200 INR
$30-250 USD
$8-15 USD / hour
₹600-1500 INR
$100-200 USD
$250-750 USD
$10-30 CAD
£250-750 GBP
$30-250 USD
₹600-1500 INR
$30-250 USD