AI Vision Specialist for Advanced Gaze Tracking -- 2
Budget: ₹750 – ₹1,250 INR
Project Title: AI Vision Expert for Camera Calibration & Gaze Tracking with Heatmap Mapping
Project Overview:
We are seeking an experienced AI specialist with a strong background in camera vision systems to assist with the development of an advanced gaze tracking solution. The goal is to accurately track a user's gaze or head direction using camera input, map the focus on the screen, and generate dynamic heatmaps representing visual attention across the screen area.
---
Key Responsibilities:
* Develop and optimize camera calibration models for accurate tracking and inference.
* Implement or integrate gaze tracking algorithms, prioritizing eye detection for precision.
* Design fallback mechanisms using head pose estimation when eye tracking is unreliable, applying a less concentrated impact point on the heatmap accordingly.
* Generate real-time or near real-time heatmaps on the screen based on user gaze/head direction and the active screen environment.
* Work with Roboflow Inference ([GitHub Link](https://github.com/roboflow/inference)) and suggest potential model integrations.
* Evaluate and propose improvements to existing workflows and systems such as gazeLLE or related libraries.
* Collaborate on calibration validation, which will be a key focus for the upcoming week.
---
Preferred Skills and Experience:
* Proven expertise in AI and computer vision, especially in gaze and head pose tracking.
* Solid understanding of camera calibration techniques and tools.
* Experience with Roboflow Inference or similar model-serving platforms.
* Strong programming skills in Python (PyTorch, OpenCV, etc.).
* Familiarity with screen interaction analysis and heatmap visualization techniques.
* Knowledge of current models or research for eye tracking and facial landmarks is a plus.
* Ability to suggest and integrate improvements in accuracy, latency, and user adaptability.
To Apply:
Please include:
* A brief overview of relevant past projects or research.
* Links to any public GitHub repositories or demos.
* Your availability over the next few days
* Any suggestions you may already have for improving this workflow.
Project Overview:
We are seeking an experienced AI specialist with a strong background in camera vision systems to assist with the development of an advanced gaze tracking solution. The goal is to accurately track a user's gaze or head direction using camera input, map the focus on the screen, and generate dynamic heatmaps representing visual attention across the screen area.
---
Key Responsibilities:
* Develop and optimize camera calibration models for accurate tracking and inference.
* Implement or integrate gaze tracking algorithms, prioritizing eye detection for precision.
* Design fallback mechanisms using head pose estimation when eye tracking is unreliable, applying a less concentrated impact point on the heatmap accordingly.
* Generate real-time or near real-time heatmaps on the screen based on user gaze/head direction and the active screen environment.
* Work with Roboflow Inference ([GitHub Link](https://github.com/roboflow/inference)) and suggest potential model integrations.
* Evaluate and propose improvements to existing workflows and systems such as gazeLLE or related libraries.
* Collaborate on calibration validation, which will be a key focus for the upcoming week.
---
Preferred Skills and Experience:
* Proven expertise in AI and computer vision, especially in gaze and head pose tracking.
* Solid understanding of camera calibration techniques and tools.
* Experience with Roboflow Inference or similar model-serving platforms.
* Strong programming skills in Python (PyTorch, OpenCV, etc.).
* Familiarity with screen interaction analysis and heatmap visualization techniques.
* Knowledge of current models or research for eye tracking and facial landmarks is a plus.
* Ability to suggest and integrate improvements in accuracy, latency, and user adaptability.
To Apply:
Please include:
* A brief overview of relevant past projects or research.
* Links to any public GitHub repositories or demos.
* Your availability over the next few days
* Any suggestions you may already have for improving this workflow.