File size: 3,616 Bytes
7a2c9ac
 
 
 
 
 
 
 
4b42627
7a2c9ac
f80cddb
 
 
 
 
 
 
 
 
 
 
7a2c9ac
f80cddb
 
 
 
 
 
 
7a2c9ac
f80cddb
 
 
1809048
f80cddb
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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

import torch
import torch.nn.functional as F
import logging
import os
import os.path as osp

import sys
CODE_SPACE=os.path.dirname(os.path.dirname(os.path.abspath(__file__)))

try:
    from mmcv.utils import Config, DictAction
except:
    from mmengine import Config, DictAction
from mono.utils.logger import setup_logger
import glob
from mono.utils.comm import init_env
from mono.model.monodepth_model import get_configured_monodepth_model
from mono.utils.running import load_ckpt
from mono.utils.do_test import transform_test_data_scalecano, get_prediction
from mono.utils.custom_data import load_from_annos, load_data

from mono.utils.avg_meter import MetricAverageMeter
from mono.utils.visualization import save_val_imgs, create_html, save_raw_imgs, save_normal_val_imgs
import cv2
from tqdm import tqdm
import numpy as np
from PIL import Image
import matplotlib.pyplot as plt

from mono.utils.unproj_pcd import reconstruct_pcd, save_point_cloud
import gradio as gr

print(CODE_SPACE, osp.exists('./mono/configs/HourglassDecoder/vit.raft5.small.py'))
cfg = Config.fromfile('./mono/configs/_base_/models/encoder_decoder/dino_vit_small_reg.dpt_raft.py')

torch.hub.download_url_to_file('https://images.unsplash.com/photo-1437622368342-7a3d73a34c8f', 'turtle.jpg')
torch.hub.download_url_to_file('https://images.unsplash.com/photo-1519066629447-267fffa62d4b', 'lions.jpg')

model = get_configured_monodepth_model(cfg, )
model, _,  _, _ = load_ckpt('./weight/metric_depth_vit_small_800k.pth', model, strict_match=False)
model.eval()

device = "cpu"
model.to(device)

def depth_normal(img):
    cv_image = np.array(img) 
    img = cv2.cvtColor(cv_image, cv2.COLOR_BGR2RGB)
    intrinsic = [1000.0, 1000.0, img.shape[1]/2, img.shape[0]/2]
    rgb_input, cam_models_stacks, pad, label_scale_factor = transform_test_data_scalecano(rgb_origin, intrinsic, cfg.data_basic)

    with torch.no_grad():
        pred_depth, pred_depth_scale, scale, output = get_prediction(
                    model = model,
                    input = rgb_input,
                    cam_model = cam_models_stacks,
                    pad_info = pad,
                    scale_info = label_scale_factor,
                    gt_depth = None,
                    normalize_scale = cfg.data_basic.depth_range[1],
                    ori_shape=[img.shape[0], img.shape[1]],
                )

        pred_normal = output['normal_out_list'][0][:, :3, :, :] 
        H, W = pred_normal.shape[2:]
        pred_normal[:, :, pad[0]:H-pad[1], pad[2]:W-pad[3]]
        
    output = pred_depth.cpu().numpy()
    formatted = (output * 255 / np.max(output)).astype('uint8')
    img = Image.fromarray(formatted)
    return img
        

inputs =  gr.inputs.Image(type='pil', label="Original Image")
depth = gr.outputs.Image(type="pil",label="Output Depth")
#normal = gr.outputs.Image(type="pil",label="Output Normal")

title = "Metric3DS"
description = "Gradio demo for Metric3DS (v2) which takes in a single image for computing metric depth and surface normal. To use it, simply upload your image, or click one of the examples to load them. Read more at the links below."
article = "<p style='text-align: center'><a href='https://arxiv.org/pdf/2307.10984.pdf'>Metric3DS: Towards Zero-shot Metric 3D and Shape Prediction from A Single Image</a> | <a href='https://github.com/YvanYin/Metric3D'>Github Repo</a></p>"

examples = [
    ["turtle.jpg"],
    ["lions.jpg"]
]

gr.Interface(depth_normal, inputs, depth, title=title, description=description, article=article, examples=examples, analytics_enabled=False).launch(enable_queue=True,cache_examples=True)