I am working on a real-time face recognition project where I have two cameras (Camera X and Camera Y) set up to capture images of people passing by. The goal is to match faces from Camera X with those captured by Camera Y and display the matched images immediately. Here is a brief overview of my setup and problem:

Cameras and Image Processing:

Camera X captures multiple images of a person.
Camera Y captures images of the same person as they pass by.
Images from Camera X are grouped and the best images are selected based on clarity and facial features (e.g., smiling, eyes open).
The best images from Camera X are then compared with images from Camera Y to find a match.
Current Approach:

I use OpenCV and Dlib for image capture and face recognition.
Captured images are processed in a batch manner and then matched.
The matched image is displayed after all images have been processed and a match is found.

this is done in parallel using threading.

There is a noticeable delay between capturing images from Camera X and displaying the matched images from Camera Y.
This delay causes the system to not show the matched images in real-time, which is critical for my project.

code structure :

CameraCapture class captures images from each camera and places them in a queue.
ImageProcessor class processes images, groups them, and finds the best images.
A separate thread is responsible for displaying the matched images.