The goal of this project is to implement a free or open-source object detection model capable of identifying key objects and their positions within an image or in each frame of a video.

The model should improve Emozo AI's capabilities in detecting and tracking objects, enhancing the accuracy of visual analysis.

Research and Selection of Object Detection Model

• Explore free and open-source object detection models (e.g., YOLO, SSD, Faster R-CNN) that are suitable for Emozo AI’s needs.
• Select a model based on accuracy, performance, and ease of integration.
• Deliverable: Model selection report with reasoning and comparison of options.

Build a Proof of Concept (the Emozo team will provide guidance, as necessary)

• Develop a basic, but functional module that implements the selected model and demonstrates the object detection functionality
• Deliverable: Functional object detection model

Testing and Evaluation

• Test the model on a diverse set of images and videos to ensure accuracy and efficiency in detecting objects.
• Evaluate the developed module for integration compatibility with Emozo's platform
• Deliverable: Test report summarizing model performance and accuracy metrics.