test_disp.py

import torch

from skimage.transform import resize as imresize
from scipy.ndimage import zoom
import numpy as np
from path import Path
import argparse
from tqdm import tqdm

from models import DispNetS, PoseExpNet


parser = argparse.ArgumentParser(description='Script for DispNet testing with corresponding groundTruth',
                                 formatter_class=argparse.ArgumentDefaultsHelpFormatter)
parser.add_argument("--pretrained-dispnet", required=True, type=str, help="pretrained DispNet path")
parser.add_argument("--pretrained-posenet", default=None, type=str, help="pretrained PoseNet path (for scale factor)")
parser.add_argument("--img-height", default=128, type=int, help="Image height")
parser.add_argument("--img-width", default=416, type=int, help="Image width")
parser.add_argument("--no-resize", action='store_true', help="no resizing is done")
parser.add_argument("--min-depth", default=1e-3)
parser.add_argument("--max-depth", default=80)

parser.add_argument("--dataset-dir", default='.', type=str, help="Dataset directory")
parser.add_argument("--dataset-list", default=None, type=str, help="Dataset list file")
parser.add_argument("--output-dir", default=None, type=str, help="Output directory for saving predictions in a big 3D numpy file")

parser.add_argument("--gt-type", default='KITTI', type=str, help="GroundTruth data type", choices=['npy', 'png', 'KITTI', 'stillbox'])
parser.add_argument("--gps", '-g', action='store_true',
                    help='if selected, will get displacement from GPS for KITTI. Otherwise, will integrate speed')
parser.add_argument("--img-exts", default=['png', 'jpg', 'bmp'], nargs='*', type=str, help="images extensions to glob")

device = torch.device("cuda") if torch.cuda.is_available() else torch.device("cpu")


@torch.no_grad()
def main():
    args = parser.parse_args()
    if args.gt_type == 'KITTI':
        from kitti_eval.depth_evaluation_utils import test_framework_KITTI as test_framework
    elif args.gt_type == 'stillbox':
        from stillbox_eval.depth_evaluation_utils import test_framework_stillbox as test_framework

    disp_net = DispNetS().to(device)
    weights = torch.load(args.pretrained_dispnet)
    disp_net.load_state_dict(weights['state_dict'])
    disp_net.eval()

    if args.pretrained_posenet is None:
        print('no PoseNet specified, scale_factor will be determined by median ratio, which is kiiinda cheating\
            (but consistent with original paper)')
        seq_length = 1
    else:
        weights = torch.load(args.pretrained_posenet)
        seq_length = int(weights['state_dict']['conv1.0.weight'].size(1)/3)
        pose_net = PoseExpNet(nb_ref_imgs=seq_length - 1, output_exp=False).to(device)
        pose_net.load_state_dict(weights['state_dict'], strict=False)

    dataset_dir = Path(args.dataset_dir)
    if args.dataset_list is not None:
        with open(args.dataset_list, 'r') as f:
            test_files = list(f.read().splitlines())
    else:
        test_files = [file.relpathto(dataset_dir) for file in sum([dataset_dir.files('*.{}'.format(ext)) for ext in args.img_exts], [])]

    framework = test_framework(dataset_dir, test_files, seq_length,
                               args.min_depth, args.max_depth,
                               use_gps=args.gps)

    print('{} files to test'.format(len(test_files)))
    errors = np.zeros((2, 9, len(test_files)), np.float32)
    if args.output_dir is not None:
        output_dir = Path(args.output_dir)
        output_dir.makedirs_p()

    for j, sample in enumerate(tqdm(framework)):
        tgt_img = sample['tgt']

        ref_imgs = sample['ref']

        h,w,_ = tgt_img.shape
        if (not args.no_resize) and (h != args.img_height or w != args.img_width):
            tgt_img = imresize(tgt_img, (args.img_height, args.img_width)).astype(np.float32)
            ref_imgs = [imresize(img, (args.img_height, args.img_width)).astype(np.float32) for img in ref_imgs]

        tgt_img = np.transpose(tgt_img, (2, 0, 1))
        ref_imgs = [np.transpose(img, (2,0,1)) for img in ref_imgs]

        tgt_img = torch.from_numpy(tgt_img).unsqueeze(0)
        tgt_img = ((tgt_img/255 - 0.5)/0.5).to(device)

        for i, img in enumerate(ref_imgs):
            img = torch.from_numpy(img).unsqueeze(0)
            img = ((img/255 - 0.5)/0.5).to(device)
            ref_imgs[i] = img

        pred_disp = disp_net(tgt_img).cpu().numpy()[0,0]

        if args.output_dir is not None:
            if j == 0:
                predictions = np.zeros((len(test_files), *pred_disp.shape))
            predictions[j] = 1/pred_disp

        gt_depth = sample['gt_depth']

        pred_depth = 1/pred_disp
        pred_depth_zoomed = zoom(pred_depth,
                                 (gt_depth.shape[0]/pred_depth.shape[0],
                                  gt_depth.shape[1]/pred_depth.shape[1])
                                 ).clip(args.min_depth, args.max_depth)
        if sample['mask'] is not None:
            pred_depth_zoomed = pred_depth_zoomed[sample['mask']]
            gt_depth = gt_depth[sample['mask']]

        if seq_length > 1:
            # Reorganize ref_imgs : tgt is middle frame but not necessarily the one used in DispNetS
            # (in case sample to test was in end or beginning of the image sequence)
            middle_index = seq_length//2
            tgt = ref_imgs[middle_index]
            reorganized_refs = ref_imgs[:middle_index] + ref_imgs[middle_index + 1:]
            _, poses = pose_net(tgt, reorganized_refs)
            displacement_magnitudes = poses[0,:,:3].norm(2,1).cpu().numpy()

            scale_factor = np.mean(sample['displacements'] / displacement_magnitudes)
            errors[0,:,j] = compute_errors(gt_depth, pred_depth_zoomed*scale_factor)

        scale_factor = np.median(gt_depth)/np.median(pred_depth_zoomed)
        errors[1,:,j] = compute_errors(gt_depth, pred_depth_zoomed*scale_factor)

    mean_errors = errors.mean(2)
    error_names = ['abs_diff', 'abs_rel','sq_rel','rms','log_rms', 'abs_log', 'a1','a2','a3']
    if args.pretrained_posenet:
        print("Results with scale factor determined by PoseNet : ")
        print("{:>10}, {:>10}, {:>10}, {:>10}, {:>10}, {:>10}, {:>10}, {:>10}, {:>10}".format(*error_names))
        print("{:10.4f}, {:10.4f}, {:10.4f}, {:10.4f}, {:10.4f}, {:10.4f}, {:10.4f}, {:10.4f}, {:10.4f}".format(*mean_errors[0]))

    print("Results with scale factor determined by GT/prediction ratio (like the original paper) : ")
    print("{:>10}, {:>10}, {:>10}, {:>10}, {:>10}, {:>10}, {:>10}, {:>10}, {:>10}".format(*error_names))
    print("{:10.4f}, {:10.4f}, {:10.4f}, {:10.4f}, {:10.4f}, {:10.4f}, {:10.4f}, {:10.4f}, {:10.4f}".format(*mean_errors[1]))

    if args.output_dir is not None:
        np.save(output_dir/'predictions.npy', predictions)


def compute_errors(gt, pred):
    thresh = np.maximum((gt / pred), (pred / gt))
    a1 = (thresh < 1.25   ).mean()
    a2 = (thresh < 1.25 ** 2).mean()
    a3 = (thresh < 1.25 ** 3).mean()

    rmse = (gt - pred) ** 2
    rmse = np.sqrt(rmse.mean())

    rmse_log = (np.log(gt) - np.log(pred)) ** 2
    rmse_log = np.sqrt(rmse_log.mean())

    abs_log = np.mean(np.abs(np.log(gt) - np.log(pred)))

    abs_rel = np.mean(np.abs(gt - pred) / gt)
    abs_diff = np.mean(np.abs(gt - pred))

    sq_rel = np.mean(((gt - pred)**2) / gt)

    return abs_diff, abs_rel, sq_rel, rmse, rmse_log, abs_log, a1, a2, a3


if __name__ == '__main__':
    main()