Detection Model
Budget: ₹1,500 – ₹12,500 INR
I am seeking a professional to develop a highly accurate dual-model system for detecting a specific object and analyzing facial expressions in both daytime and nighttime images. The system should be able to:
- Object Detection Model: Identify the target object accurately without confusing other individuals (e.g., people nearby), even when multiple entities are in the frame.
- Draw a bounding box around the target object for focused detection.
- Apply masking to eliminate everything outside the bounding box for privacy.
- Facial Expression Analysis Model: Detect and interpret the facial expressions of the target to assess emotional or physical states (e.g., discomfort, happiness).
This model should work in various lighting conditions and be optimized for use on an embedded device.
Deployment: Both models will be deployed on an ESP32-CAM, so they must be optimized for embedded devices. Deliverables should include a sketch file, TFLite files for each model, and complete embedded logic code.
Data Requirements: Currently, there is no labeled training data available, so the project may involve data collection and annotation to ensure high accuracy for both models.
Expected Accuracy: Both models should achieve a high level of accuracy across various lighting conditions (day and night) to ensure reliable detection and analysis.
Deliverables:
-Complete codebase
-Dataset used (including labeled images for both object detection and facial expressions)
-TFLite files for each model
-C Array files for each model
-Sketch file with all embedded logic
Ideal Skills and Experience:
Machine Learning & Computer Vision: Proven expertise in developing detection and expression analysis models, especially for edge devices.
Embedded Systems: Experience with ESP32-CAM, TinyML, and model optimization.
Data Collection & Annotation: Ability to source or capture relevant images, label data, and create an effective training dataset.
Track Record: Demonstrated success in building high-accuracy detection and expression analysis models.
- Object Detection Model: Identify the target object accurately without confusing other individuals (e.g., people nearby), even when multiple entities are in the frame.
- Draw a bounding box around the target object for focused detection.
- Apply masking to eliminate everything outside the bounding box for privacy.
- Facial Expression Analysis Model: Detect and interpret the facial expressions of the target to assess emotional or physical states (e.g., discomfort, happiness).
This model should work in various lighting conditions and be optimized for use on an embedded device.
Deployment: Both models will be deployed on an ESP32-CAM, so they must be optimized for embedded devices. Deliverables should include a sketch file, TFLite files for each model, and complete embedded logic code.
Data Requirements: Currently, there is no labeled training data available, so the project may involve data collection and annotation to ensure high accuracy for both models.
Expected Accuracy: Both models should achieve a high level of accuracy across various lighting conditions (day and night) to ensure reliable detection and analysis.
Deliverables:
-Complete codebase
-Dataset used (including labeled images for both object detection and facial expressions)
-TFLite files for each model
-C Array files for each model
-Sketch file with all embedded logic
Ideal Skills and Experience:
Machine Learning & Computer Vision: Proven expertise in developing detection and expression analysis models, especially for edge devices.
Embedded Systems: Experience with ESP32-CAM, TinyML, and model optimization.
Data Collection & Annotation: Ability to source or capture relevant images, label data, and create an effective training dataset.
Track Record: Demonstrated success in building high-accuracy detection and expression analysis models.