update dataset and network

.gitignore

@@ -0,0 +1,3 @@
+*.pyc
+*.png
+*.jpg

dataset/cityscapes_dataset.py

@@ -0,0 +1,128 @@
+import os
+import os.path as osp
+import json
+import numpy as np
+import random
+import matplotlib.pyplot as plt
+import collections
+import torch
+import torchvision
+from torch.utils import data
+from PIL import Image
+from .utils import RandomResizedCrop
+
+
+def label_mapping(input, mapping):
+    output = np.copy(input)
+    for ind in range(len(mapping)):
+        output[input == mapping[ind][0]] = mapping[ind][1]
+    return np.array(output, dtype=np.float32)
+
+
+class cityscapesDataSet(data.Dataset):
+    def __init__(self, root, list_path, max_iters=None, crop_size=(1024, 512), ignore_label=255, transform = None, set='val', dataset_info = None, need_label = False):
+        self.root = root
+        self.list_path = list_path
+        self.crop_size = crop_size
+        self.ignore_label = ignore_label
+        # self.mean_bgr = np.array([104.00698793, 116.66876762, 122.67891434])
+        self.img_ids = [i_id.strip() for i_id in open(list_path)]
+        if not max_iters==None:
+            self.img_ids = self.img_ids * int(np.ceil(float(max_iters) / len(self.img_ids)))
+        self.files = []
+        self.set = set
+        self.transform = transform
+        self.need_label = need_label
+        with open('./dataset/cityscapes_list/info.json', 'r') as fp:
+            dataset_info = json.load(fp)
+        self.mapping = np.array(dataset_info['label2train'], dtype=np.int)
+
+        # for split in ["train", "trainval", "val"]:
+        for name in self.img_ids:
+            name = name.split(" ")[0]
+            img_file = osp.join(self.root, "leftImg8bit/%s/%s" % (self.set, name))
+            lbl_name = "_".join(name.split("_")[:-1]) + "_gtFine_labelIds.png"
+            lbl_file = osp.join(self.root, "gtFine/%s/%s"%(self.set, lbl_name))
+            self.files.append({
+                "img": img_file,
+                "name": name,
+                "label": lbl_file
+            })
+
+    def __len__(self):
+        return len(self.files)
+
+    def __getitem__(self, index):
+        datafiles = self.files[index]
+
+        image = Image.open(datafiles["img"]).convert('RGB')
+        name = datafiles["name"]
+
+        if self.need_label == True:
+
+            label = Image.open(datafiles["label"])
+            label = label_mapping(np.asarray(label), self.mapping)
+            label = Image.fromarray(label)
+            if self.transform:
+                image, label = self.transform(image, label)
+        else:
+            if self.transform:
+                image = self.transform(image)
+
+
+        if self.need_label == True:
+
+            return image, label, name
+        else:
+            return image, name
+
+
+
+class fake_cityscapesDataSet(data.Dataset):
+    def __init__(self, root, list_path, max_iters=None, crop_size=(1024, 512), ignore_label=255, transform = None, set='val', dataset_info = None, need_label = False):
+        self.root = root
+        self.list_path = list_path
+        self.crop_size = crop_size
+        self.ignore_label = ignore_label
+        # self.mean_bgr = np.array([104.00698793, 116.66876762, 122.67891434])
+        self.img_ids = [i_id.strip() for i_id in open(list_path)]
+        if not max_iters==None:
+            self.img_ids = self.img_ids * int(np.ceil(float(max_iters) / len(self.img_ids)))
+        self.files = []
+        self.set = set
+        self.transform = transform
+        self.need_label = True
+        with open('./dataset/cityscapes_list/info.json', 'r') as fp:
+           dataset_info = json.load(fp)
+        self.mapping = np.array(dataset_info['label2train'], dtype=np.int)
+
+        # for split in ["train", "trainval", "val"]:
+        for name in self.img_ids:
+            name = name.split(" ")[0]
+            img_file = osp.join(self.root, "leftImg8bit/%s/%s" % (self.set, name))
+            lbl_name = "_".join(name.split("_")[:-1]) + "_gtFine_labelIds.png"
+            lbl_file = osp.join(self.root, "gtFine/%s/%s"%(self.set, lbl_name))
+            self.files.append({
+                "img": img_file,
+                "name": name,
+                "label": lbl_file
+            })
+
+    def __len__(self):
+        return len(self.files)
+
+    def __getitem__(self, index):
+        datafiles = self.files[index]
+
+        image = Image.open(datafiles["img"]).convert('RGB')
+        name = datafiles["name"]
+
+        if self.need_label == True:
+
+            label = np.zeros((560, 480), dtype=np.uint8)
+            label = Image.fromarray(label)
+            if self.transform:
+                image, label = self.transform(image, label)
+
+        return image, label, name
+

dataset/cityscapes_list/info.json

@@ -0,0 +1,88 @@
+{
+  "classes":19,
+  "label2train":[
+    [0, 255],
+    [1, 255],
+    [2, 255],
+    [3, 255],
+    [4, 255],
+    [5, 255],
+    [6, 255],
+    [7, 0],
+    [8, 1],
+    [9, 255],
+    [10, 255],
+    [11, 2],
+    [12, 3],
+    [13, 4],
+    [14, 255],
+    [15, 255],
+    [16, 255],
+    [17, 5],
+    [18, 255],
+    [19, 6],
+    [20, 7],
+    [21, 8],
+    [22, 9],
+    [23, 10],
+    [24, 11],
+    [25, 12],
+    [26, 13],
+    [27, 14],
+    [28, 15],
+    [29, 255],
+    [30, 255],
+    [31, 16],
+    [32, 17],
+    [33, 18],
+    [-1, 255]],
+  "label":[
+    "road",
+    "sidewalk",
+    "building",
+    "wall",
+    "fence",
+    "pole",
+    "light",
+    "sign",
+    "vegetation",
+    "terrain",
+    "sky",
+    "person",
+    "rider",
+    "car",
+    "truck",
+    "bus",
+    "train",
+    "motocycle",
+    "bicycle"],
+  "palette":[
+    [128,64,128],
+    [244,35,232],
+    [70,70,70],
+    [102,102,156],
+    [190,153,153],
+    [153,153,153],
+    [250,170,30],
+    [220,220,0],
+    [107,142,35],
+    [152,251,152],
+    [70,130,180],
+    [220,20,60],
+    [255,0,0],
+    [0,0,142],
+    [0,0,70],
+    [0,60,100],
+    [0,80,100],
+    [0,0,230],
+    [119,11,32],
+    [0,0,0]],
+  "mean":[
+    73.158359210711552,
+    82.908917542625858,
+    72.392398761941593],
+  "std":[
+    47.675755341814678,
+    48.494214368814916,
+    47.736546325441594]
+}