Name		Name	Last commit message	Last commit date
parent directory ..
LICENSE		LICENSE
README.md		README.md
input.jpg		input.jpg
input_depth.png		input_depth.png
midas.py		midas.py

README.md

MiDaS

Input

(Image from kitti dataset http://www.cvlibs.net/datasets/kitti/raw_data.php)

Shape : (1, h, w, 3)

Output

Shape : (1, h, w)

Usage

Automatically downloads the tflite file on the first run. It is necessary to be connected to the Internet while downloading.

For the sample image,

$ python3 midas.py

If you want to specify the input image, put the image path after the --input option.
You can use --savepath option to change the name of the output file to save.

$ python3 midas.py --input IMAGE_PATH --savepath SAVE_IMAGE_PATH

By adding the --video option, you can input the video.
If you pass 0 as an argument to VIDEO_PATH, you can use the webcam input instead of the video file.

$ python3 midas.py --video VIDEO_PATH

By adding the -v21 option, you can use version 2.1 model.
(default use version 2.0 model)

If you use the version 2.1 model, you can use the small model with the --model_type small option.

$ python3 midas.py -v21 --model_type small

Two versions of the model are provided: full integer quantization (8-bit) and full precision floating point (32-bit). By default, the full integer quantization is used but the user can select the other version by passing the --float flag.