Proxy-GS/metrics_mask.py at main · Visionary-Laboratory/Proxy-GS · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
#
# Copyright (C) 2023, Inria
# GRAPHDECO research group, https://team.inria.fr/graphdeco
# All rights reserved.
#
# This software is free for non-commercial, research and evaluation use
# under the terms of the LICENSE.md file.
#
# For inquiries contact  george.drettakis@inria.fr
#

from pathlib import Path
import os
from PIL import Image
import torch
import torchvision.transforms.functional as tf
from utils.loss_utils import ssim
# from lpipsPyTorch import lpips
import lpips
import json
from tqdm import tqdm
from utils.image_utils import psnr
from argparse import ArgumentParser
import numpy as np
def PILtoTorch(pil_image, resolution):
    resized_image_PIL = pil_image.resize(resolution)
    resized_image = torch.from_numpy(np.array(resized_image_PIL)) / 255.0
    if len(resized_image.shape) == 3:
        return resized_image.permute(2, 0, 1)
    else:
        return resized_image.unsqueeze(dim=-1).permute(2, 0, 1)
def readImages(renders_dir, gt_dir, mask_dir):
    renders = []
    gts = []
    image_names = []
    masks = []
    for fname in os.listdir(renders_dir):
        render = Image.open(renders_dir / fname)
        gt = Image.open(gt_dir / fname)
        mask_path = mask_dir[0]+'/'+fname
        mask = Image.open(mask_path)
        orig_w, orig_h = gt.size
        resolution = (orig_w, orig_h)
        new_mask = PILtoTorch(mask, resolution)

        renders.append(tf.to_tensor(render).unsqueeze(0)[:, :3, :, :].cuda())
        gts.append(tf.to_tensor(gt).unsqueeze(0)[:, :3, :, :].cuda())
        masks.append(new_mask.cuda())
        image_names.append(fname)
    return renders, gts, image_names, masks

def evaluate(model_paths, mask_paths):

    full_dict = {}
    per_view_dict = {}
    full_dict_polytopeonly = {}
    per_view_dict_polytopeonly = {}
    print("")

    for scene_dir in model_paths:
        print("Scene:", scene_dir)
        full_dict[scene_dir] = {}
        per_view_dict[scene_dir] = {}
        full_dict_polytopeonly[scene_dir] = {}
        per_view_dict_polytopeonly[scene_dir] = {}

        test_dir = Path(scene_dir) / "test"

        for method in os.listdir(test_dir):
            print("Method:", method)

            full_dict[scene_dir][method] = {}
            per_view_dict[scene_dir][method] = {}
            full_dict_polytopeonly[scene_dir][method] = {}
            per_view_dict_polytopeonly[scene_dir][method] = {}

            method_dir = test_dir / method
            gt_dir = method_dir/ "gt"
            renders_dir = method_dir / "renders"
            json_path = method_dir / "per_view_count.json"
            renders, gts, image_names, masks = readImages(renders_dir, gt_dir, mask_paths)

            json_file = open(json_path)
            gs_data = json.load(json_file)
            json_file.close()
            ssims = []
            psnrs = []
            lpipss = []
            gss = []


            for idx in tqdm(range(len(renders)), desc="Metric evaluation progress"):

                image_name = "{:05d}.png".format(idx)
                gss.append(gs_data[image_name])
                ssims.append(ssim(renders[idx] * masks[idx], gts[idx] * masks[idx]))
                psnrs.append(psnr(renders[idx] * masks[idx], gts[idx] * masks[idx]))
                lpipss.append(lpips_fn(renders[idx] * masks[idx], gts[idx] * masks[idx]).detach())


            print("  SSIM : {:>12.7f}".format(torch.tensor(ssims).mean(), ".5"))
            print("  PSNR : {:>12.7f}".format(torch.tensor(psnrs).mean(), ".5"))
            print("  LPIPS: {:>12.7f}".format(torch.tensor(lpipss).mean(), ".5"))
            print("  GS   : {:>12.7f}".format(torch.tensor(gss).float().mean(), ".5"))
            print("")

            full_dict[scene_dir][method].update({"SSIM": torch.tensor(ssims).mean().item(),
                                                    "PSNR": torch.tensor(psnrs).mean().item(),
                                                    "LPIPS": torch.tensor(lpipss).mean().item(),
                                                    "GS": torch.tensor(gss).float().mean().item()})
            per_view_dict[scene_dir][method].update({"SSIM": {name: ssim for ssim, name in zip(torch.tensor(ssims).tolist(), image_names)},
                                                        "PSNR": {name: psnr for psnr, name in zip(torch.tensor(psnrs).tolist(), image_names)},
                                                        "LPIPS": {name: lp for lp, name in zip(torch.tensor(lpipss).tolist(), image_names)}})

        with open(scene_dir + "/results.json", 'w') as fp:
            json.dump(full_dict[scene_dir], fp, indent=True)
        with open(scene_dir + "/per_view.json", 'w') as fp:
            json.dump(per_view_dict[scene_dir], fp, indent=True)

if __name__ == "__main__":
    device = torch.device("cuda:0")
    torch.cuda.set_device(device)
    lpips_fn = lpips.LPIPS(net='vgg').to(device)

    # Set up command line argument parser
    parser = ArgumentParser(description="Training script parameters")
    parser.add_argument('--model_paths', '-m', required=True, nargs="+", type=str, default=[])
    parser.add_argument('--mask_paths', '-s', required=True, nargs="+", type=str, default=[])
    args = parser.parse_args()
    evaluate(args.model_paths, args.mask_paths)