Model description

The original idea from Keras examples Monocular depth estimation of author Victor Basu

Full credits go to Vu Minh Chien

Depth estimation is a crucial step towards inferring scene geometry from 2D images. The goal in monocular depth estimation is to predict the depth value of each pixel or infer depth information, given only a single RGB image as input.

Dataset

NYU Depth Dataset V2 is comprised of video sequences from a variety of indoor scenes as recorded by both the RGB and Depth cameras from the Microsoft Kinect.

Training procedure

Training hyperparameters

Model architecture :

• UNet with a pretrained DenseNet 201 backbone.

The following hyperparameters were used during training:

• learning_rate: 1e-04
• train_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: ReduceLROnPlateau
• num_epochs: 10

Training results

View Model Demo