HR-Depth: High Resolution Self-Supervised Monocular Depth Estimation

This repository contains an executable demonstration of ONNX and TensorFlow implementation for depth estimation using the model proposed in:

HR-Depth: High Resolution Self-Supervised Monocular Depth Estimation Xiaoyang Lyu, Liang Liu, Mengmeng Wang, Xin Kong, Lina Liu, Yong Liu*, Xinxin Chen and Yi Yuan

official implementation repository

1. Requirements Installation

Inside the directory containing the cloned repo, install the necessary packages in requirements.txt

pip install -r requirements.txt

2. Downloading Trained Models

All of the trained models can be downloaded by running the following three lines of code in terminal:

for f in *.sh; do
  bash "$f" 
done

Alternatively, only the shell script of the eponymous model of interest can be executed.

3.A. HR-Depth with ONNX Runtime in Python

To prepare the downloaded ONNX models, execute the following command:

python onnx_merge.py

the following output should be printed to console upon successful execution

graph1 outputs: ['317a', '852a', '870a', '897a', '836a']
graph2 inputs: ['0b', 'input.1b', 'input.13b', 'input.25b', 'input.37b']
Constructing the io_match list from your input and output

To run a demo of the HR-Depth with ONNX, execute the following command:

python demo/demo_hr_depth_onnx.py

3.B. HR-Depth with TensorFlow Lite in Python

To run a demo of the HR-Depth with TensorFlow, execute the following command:

python demo_hr_depth_tflite.py

Written with StackEdit.