init

Author: jiaxue-ai

README.md

@@ -0,0 +1,27 @@
+# Pytorch Material Classification
+
+
+## Setup
+
+### Prerequisites
+
+- Ubuntu 18.04
+- Pytorch 1.1.0
+
+### Getting Started
+
+- Clone this repo:
+```bash
+git clone git@github.com:jiaxue1993/pytorch-material-classification.git
+cd pytorch-material-classification/
+```
+- Install [Pytorch-Encoding](https://github.com/zhanghang1989/PyTorch-Encoding) 
+- Download the [database](https://drive.google.com/file/d/1Hd1G7aKhsPPMbNrk4zHNJAzoXvUzWJ9M/view?usp=sharing), release to dataset folder
+- Run the code by bash run.sh
+
+## Datasets
+You can find the database [here](http://jiaxueweb.com/)
+
+## Acknowledgement
+
+This work was supported by National Science Foundation award IIS-1421134. A GPU used for this research was donated by the NVIDIA Corporation.

dataloader/__init__.py

@@ -0,0 +1,49 @@
+import torch
+import torch.utils.data as data
+import torchvision
+from torchvision import transforms, datasets
+
+from PIL import Image
+import os
+import os.path
+
+
+def find_classes(dir):
+    classes = [d for d in os.listdir(dir) if os.path.isdir(os.path.join(dir, d))]
+    classes.sort()
+    class_to_idx = {classes[i]: i for i in range(len(classes))}
+    return classes, class_to_idx
+
+
+class Dataloder():
+    def __init__(self, config):
+        normalize = transforms.Normalize(mean=[0.485, 0.456, 0.406],
+                                         std=[0.229, 0.224, 0.225])
+        transform_train = transforms.Compose([
+            transforms.Resize(256),
+            transforms.RandomResizedCrop(224),
+            transforms.RandomHorizontalFlip(),
+            transforms.ToTensor(),
+            normalize,
+        ])
+        transform_test = transforms.Compose([
+            transforms.Resize(256),
+            transforms.CenterCrop(224),
+            transforms.ToTensor(),
+            normalize,
+        ])
+
+        trainset = datasets.ImageFolder(os.path.join(config.dataset_path, 'train'), transform_train)
+        testset = datasets.ImageFolder(os.path.join(config.dataset_path, 'test'), transform_test)
+    
+        kwargs = {'num_workers': 8, 'pin_memory': True}
+        trainloader = torch.utils.data.DataLoader(trainset, batch_size=
+            config.batch_size, shuffle=True, **kwargs)
+        testloader = torch.utils.data.DataLoader(testset, batch_size=
+            config.batch_size, shuffle=False, **kwargs)
+        self.classes = trainset.classes
+        self.trainloader = trainloader 
+        self.testloader = testloader
+    
+    def getloader(self):
+        return self.trainloader, self.testloader

dataloader/dtd.py

@@ -0,0 +1,86 @@
+import torch
+import torch.utils.data as data
+import torchvision
+from torchvision import datasets, models, transforms
+
+from PIL import Image
+import os
+import os.path
+
+_imagenet_pca = {
+    'eigval': torch.Tensor([0.2175, 0.0188, 0.0045]),
+    'eigvec': torch.Tensor([
+        [-0.5675,  0.7192,  0.4009],
+        [-0.5808, -0.0045, -0.8140],
+        [-0.5836, -0.6948,  0.4203],
+    ])
+}
+
+
+class Dataloder():
+    def __init__(self, config):
+        normalize = transforms.Normalize(mean=[0.485, 0.456, 0.406],
+                                         std=[0.229, 0.224, 0.225])
+        transform_train = transforms.Compose([
+            transforms.Resize(256),
+            transforms.RandomResizedCrop(224),
+            transforms.RandomHorizontalFlip(),
+            transforms.ColorJitter(0.4,0.4,0.4),
+            transforms.ToTensor(),
+            Lighting(0.1, _imagenet_pca['eigval'], _imagenet_pca['eigvec']),
+            normalize,
+        ])
+        transform_test = transforms.Compose([
+            transforms.Resize(256),
+            transforms.CenterCrop(224),
+            transforms.ToTensor(),
+            normalize,
+        ])
+
+        data_dir = 'dataset/gtos-mobile'
+        trainset = datasets.ImageFolder(os.path.join(data_dir, 'train'), transform_train)
+        testset = datasets.ImageFolder(os.path.join(data_dir, 'test'), transform_test)
+
+
+        kwargs = {'num_workers': 8, 'pin_memory': True} if config.cuda else {}
+        trainloader = torch.utils.data.DataLoader(trainset, batch_size=
+            config.batch_size, shuffle=True, **kwargs)
+        testloader = torch.utils.data.DataLoader(testset, batch_size=
+            config.batch_size, shuffle=False, **kwargs)
+
+
+        self.trainloader = trainloader 
+        self.testloader = testloader
+        self.classes = trainset.classes
+    
+    def getloader(self):
+        return self.classes, self.trainloader, self.testloader
+
+
+class Lighting(object):
+    """Lighting noise(AlexNet - style PCA - based noise)"""
+
+    def __init__(self, alphastd, eigval, eigvec):
+        self.alphastd = alphastd
+        self.eigval = eigval
+        self.eigvec = eigvec
+
+    def __call__(self, img):
+        if self.alphastd == 0:
+            return img
+
+        alpha = img.new().resize_(3).normal_(0, self.alphastd)
+        rgb = self.eigvec.type_as(img).clone()\
+            .mul(alpha.view(1, 3).expand(3, 3))\
+            .mul(self.eigval.view(1, 3).expand(3, 3))\
+            .sum(1).squeeze()
+
+        return img.add(rgb.view(3, 1, 1).expand_as(img))
+
+
+if __name__ == "__main__":
+    data_dir = 'dataset/gtos-mobile'
+    trainset = datasets.ImageFolder(os.path.join(data_dir, 'train'))
+    testset = datasets.ImageFolder(os.path.join(data_dir, 'test'))
+    print(trainset.classes)
+    print(len(testset))

dataloader/gtos.py

@@ -0,0 +1,88 @@
+import torch
+import torch.utils.data as data
+import torchvision
+from torchvision import transforms
+
+from PIL import Image
+import os
+import os.path
+
+
+def find_classes(dir):
+    classes = [d for d in os.listdir(dir) if os.path.isdir(os.path.join(dir, d))]
+    classes.sort()
+    class_to_idx = {classes[i]: i for i in range(len(classes))}
+    return classes, class_to_idx
+
+
+def make_dataset(filename, datadir, class_to_idx):
+    images = []
+    labels = []
+    with open(os.path.join(filename), "r") as lines:
+        for line in lines:
+            _image = os.path.join(datadir, line.rstrip('\n'))
+            _dirname = os.path.split(os.path.dirname(_image))[1]
+            assert os.path.isfile(_image)
+            label = class_to_idx[_dirname]
+            images.append(_image)
+            labels.append(label)
+
+    return images, labels
+
+
+class MINCDataloder(data.Dataset):
+    def __init__(self, root, filename, transform=None):
+        classes, class_to_idx = find_classes(root + '/images')
+        self.classes = classes
+        self.class_to_idx = class_to_idx
+        self.transform = transform
+        self.images, self.labels = make_dataset(filename, root, 
+            class_to_idx)
+        assert (len(self.images) == len(self.labels))
+
+    def __getitem__(self, index):
+        _img = Image.open(self.images[index]).convert('RGB')
+        _label = self.labels[index]
+        if self.transform is not None:
+            _img = self.transform(_img)
+
+        return _img, _label
+
+    def __len__(self):
+        return len(self.images)
+
+
+class Dataloder():
+    def __init__(self, config):
+        normalize = transforms.Normalize(mean=[0.485, 0.456, 0.406],
+                                         std=[0.229, 0.224, 0.225])
+        transform_train = transforms.Compose([
+            transforms.Resize(256),
+            transforms.RandomResizedCrop(224),
+            transforms.RandomHorizontalFlip(),
+            transforms.ToTensor(),
+            normalize,
+        ])
+        transform_test = transforms.Compose([
+            transforms.Resize(256),
+            transforms.CenterCrop(224),
+            transforms.ToTensor(),
+            normalize,
+        ])
+
+        trainset = MINCDataloder(config.dataset_path, 
+            config.train_source, transform=transform_train)
+        testset = MINCDataloder(config.dataset_path, 
+            config.eval_source, transform=transform_test)
+
+        kwargs = {'num_workers': 0, 'pin_memory': True}
+        trainloader = torch.utils.data.DataLoader(trainset, batch_size=
+            config.batch_size, shuffle=True, **kwargs)
+        testloader = torch.utils.data.DataLoader(testset, batch_size=
+            config.batch_size, shuffle=False, **kwargs)
+        self.classes = trainset.classes
+        self.trainloader = trainloader 
+        self.testloader = testloader
+    
+    def getloader(self):
+        return self.trainloader, self.testloader

dataloader/minc.py

@@ -0,0 +1,59 @@
+import os
+import os.path as osp
+import sys
+import time
+import numpy as np
+from easydict import EasyDict as edict
+import argparse
+
+os.environ['CUDA_VISIBLE_DEVICES'] = '0'
+
+C = edict()
+config = C
+
+C.seed = 0
+
+"""please config ROOT_dir and user when u first using"""
+C.repo_name = 'Deep-Encoding-Pooling-Network-DEP-'
+C.dataset = 'GTOS'
+C.model = 'DEP'
+C.abs_dir = osp.realpath(".")
+C.this_dir = C.abs_dir.split(osp.sep)[-1]
+C.root_dir = C.abs_dir[:C.abs_dir.index(C.repo_name) + len(C.repo_name)]
+C.log_dir = osp.abspath(osp.join(C.root_dir, 'log', C.this_dir))
+C.snapshot_dir = osp.abspath(osp.join(C.log_dir, "snapshot"))
+
+exp_time = time.strftime('%Y_%m_%d_%H_%M_%S', time.localtime())
+C.log_file = C.log_dir + '/log_' + exp_time + '.log'
+C.link_log_file = C.log_file + '/log_last.log'
+C.val_log_file = C.log_dir + '/val_' + exp_time + '.log'
+C.link_val_log_file = C.log_dir + '/val_last.log'
+
+"""Data Dir and Weight Dir"""
+C.dataset_path = "/home/jia/Downloads/database/minc-2500/"
+C.img_root_folder = C.dataset_path
+C.gt_root_folder = C.dataset_path
+
+
+"""Path Config"""
+
+def add_path(path):
+    if path not in sys.path:
+        sys.path.insert(0, path)
+
+
+add_path(C.root_dir)
+
+"""Train Config"""
+C.lr = 1e-2
+C.lr_decay = 40
+C.momentum = 0.9
+C.weight_decay = 1e-4
+C.batch_size = 64
+C.start_epoch = 1
+C.nepochs = 100
+
+C.no_cuda = False
+C.resume = False
+C.momentum = 0.9
+C.weight_decay = 1e-4