forward.py

import argparse
import os
import random
import time
from os.path import isfile, join, split

import torch
import torchvision
import torch.backends.cudnn as cudnn
import torch.nn as nn
import torch.optim
import numpy as np
import tqdm
import yaml
import cv2

from torch.optim import lr_scheduler
from logger import Logger

from dataloader import get_loader
from model.network import Net
from skimage.measure import label, regionprops
from utils import reverse_mapping, visulize_mapping, edge_align, get_boundary_point

parser = argparse.ArgumentParser(description='PyTorch Semantic-Line Training')
# arguments from command line
parser.add_argument('--config', default="./config.yml", help="path to config file")
parser.add_argument('--model', required=True, help='path to the pretrained model')
parser.add_argument('--align', default=False, action='store_true')
parser.add_argument('--tmp', default="", help='tmp')
args = parser.parse_args()

assert os.path.isfile(args.config)
CONFIGS = yaml.load(open(args.config))

# merge configs
if args.tmp != "" and args.tmp != CONFIGS["MISC"]["TMP"]:
    CONFIGS["MISC"]["TMP"] = args.tmp

os.makedirs(CONFIGS["MISC"]["TMP"], exist_ok=True)
logger = Logger(os.path.join(CONFIGS["MISC"]["TMP"], "log.txt"))

def main():

    logger.info(args)

    model = Net(numAngle=CONFIGS["MODEL"]["NUMANGLE"], numRho=CONFIGS["MODEL"]["NUMRHO"], backbone=CONFIGS["MODEL"]["BACKBONE"])
    model = model.cuda(device=CONFIGS["TRAIN"]["GPU_ID"])

    if args.model:
        if isfile(args.model):
            logger.info("=> loading pretrained model '{}'".format(args.model))
            checkpoint = torch.load(args.model)
            if 'state_dict' in checkpoint.keys():
                model.load_state_dict(checkpoint['state_dict'])
            else:
                model.load_state_dict(checkpoint)
            logger.info("=> loaded checkpoint '{}'"
                  .format(args.model))
        else:
            logger.info("=> no pretrained model found at '{}'".format(args.model))
    # dataloader
    test_loader = get_loader(CONFIGS["DATA"]["TEST_DIR"], CONFIGS["DATA"]["TEST_LABEL_FILE"], 
                                batch_size=1, num_thread=CONFIGS["DATA"]["WORKERS"], test=True)

    logger.info("Data loading done.")

    
    
    logger.info("Start testing.")
    total_time = test(test_loader, model, args)
    
    logger.info("Test done! Total %d imgs at %.4f secs without image io, fps: %.3f" % (len(test_loader), total_time, len(test_loader) / total_time))

        
def test(test_loader, model, args):
    # switch to evaluate mode
    model.eval()
    with torch.no_grad():
        bar = tqdm.tqdm(test_loader)
        iter_num = len(test_loader.dataset)
        ftime = 0
        ntime = 0
        for i, data in enumerate(bar):
            t = time.time()
            images, names, size = data
            
            images = images.cuda(device=CONFIGS["TRAIN"]["GPU_ID"])
            # size = (size[0].item(), size[1].item())       
            key_points = model(images)
            
            key_points = torch.sigmoid(key_points)
            ftime += (time.time() - t)
            t = time.time()
            visualize_save_path = os.path.join(CONFIGS["MISC"]["TMP"], 'visualize_test')
            os.makedirs(visualize_save_path, exist_ok=True)

            binary_kmap = key_points.squeeze().cpu().numpy() > CONFIGS['MODEL']['THRESHOLD']
            kmap_label = label(binary_kmap, connectivity=1)
            props = regionprops(kmap_label)
            plist = []
            for prop in props:
                plist.append(prop.centroid)

            size = (size[0][0], size[0][1])
            b_points = reverse_mapping(plist, numAngle=CONFIGS["MODEL"]["NUMANGLE"], numRho=CONFIGS["MODEL"]["NUMRHO"], size=(400, 400))
            scale_w = size[1] / 400
            scale_h = size[0] / 400
            for i in range(len(b_points)):
                y1 = int(np.round(b_points[i][0] * scale_h))
                x1 = int(np.round(b_points[i][1] * scale_w))
                y2 = int(np.round(b_points[i][2] * scale_h))
                x2 = int(np.round(b_points[i][3] * scale_w))
                if x1 == x2:
                    angle = -np.pi / 2
                else:
                    angle = np.arctan((y1-y2) / (x1-x2))
                (x1, y1), (x2, y2) = get_boundary_point(y1, x1, angle, size[0], size[1])
                b_points[i] = (y1, x1, y2, x2)

            vis = visulize_mapping(b_points, size[::-1], names[0])

            

            cv2.imwrite(join(visualize_save_path, names[0].split('/')[-1]), vis)
            np_data = np.array(b_points)
            np.save(join(visualize_save_path, names[0].split('/')[-1].split('.')[0]), np_data)

            if CONFIGS["MODEL"]["EDGE_ALIGN"] and args.align:
                for i in range(len(b_points)):
                    b_points[i] = edge_align(b_points[i], names[0], size, division=5)
                vis = visulize_mapping(b_points, size, names[0])
                cv2.imwrite(join(visualize_save_path, names[0].split('/')[-1].split('.')[0]+'_align.png'), vis)
                np_data = np.array(b_points)
                np.save(join(visualize_save_path, names[0].split('/')[-1].split('.')[0]+'_align'), np_data)
            ntime += (time.time() - t)
    print('forward time for total images: %.6f' % ftime)
    print('post-processing time for total images: %.6f' % ntime)
    return ftime + ntime

if __name__ == '__main__':
    main()