shariqfarooq123 / AdaBins

Official implementation of Adabins: Depth Estimation using adaptive bins
GNU General Public License v3.0
732 stars 157 forks source link

visualization of nyud predicted depth #14

Closed TianhaoFu closed 3 years ago

TianhaoFu commented 3 years ago

Hi, Thanks for your code.

Could you please provide the visualization code of nyud predicted depth or the heat map of nyud predicted depth?

Thank you very much!

shariqfarooq123 commented 3 years ago

Are you looking for this? https://github.com/shariqfarooq123/AdaBins/blob/2fb686a66a304f0a719bc53d77412460af97fd61/utils.py#L49 :

For numpy arrays:

import matplotlib
def colorize(arr, vmin=0.1, vmax=10, cmap='magma_r', ignore=-1):
    invalid_mask = arr == ignore

    # normalize
    vmin = arr.min() if vmin is None else vmin
    vmax = arr.max() if vmax is None else vmax
    if vmin != vmax:
        arr = (arr - vmin) / (vmax - vmin)  # vmin..vmax
    else:
        # Avoid 0-division
        arr = arr * 0.
    cmapper = matplotlib.cm.get_cmap(cmap)
    arr = cmapper(arr, bytes=True)  # (nxmx4)
    arr[invalid_mask] = 255
    img = arr[:, :, :3]

    return img

Cmaps used in paper are 'jet' and 'magma_r'. You can find other color maps here