3D Bounding Boxes without using neural networks

Question

The problem is that:

1. I have input ".jpg" format pictures in one folder about 1200 pictures

2. Another folder containing ".txt" flies for labels for each pictures [class_index, BB_x, BB_y, BB_w, BB_h, BB3D_x1, BB3D_x2, BB3D_y1, BB3D_y2, BB3D_z1, BB3D_z2]. Below, you can find an example using these values in attachment. Labels screenshot

3. And camera matrix as below:

   cam_mtx = np.array([ [358.5, 0.0, 512.0], [0.0, 358.5, 256.0], [0.0, 0.0, 1.0], ])

I need to develop a model to get 3d bounding boxes without using neural network techniques, any help, please.