From 8438bfe5abe67bf37c69cdc8e0bf6bbc41e39237 Mon Sep 17 00:00:00 2001
From: Nataniel Ruiz <nruiz9@gatech.edu>
Date: 星期一, 04 三月 2019 08:14:39 +0800
Subject: [PATCH] Update README.md
---
code/datasets.py | 292 +++++++++++++++++++---------------------------------------
1 files changed, 95 insertions(+), 197 deletions(-)
diff --git a/code/datasets.py b/code/datasets.py
index aee5246..e8ab9f4 100644
--- a/code/datasets.py
+++ b/code/datasets.py
@@ -1,18 +1,84 @@
-import numpy as np
-import torch
-import cv2
-from torch.utils.data.dataset import Dataset
import os
+import numpy as np
+import cv2
+import pandas as pd
+
+import torch
+from torch.utils.data.dataset import Dataset
+from torchvision import transforms
+
from PIL import Image, ImageFilter
import utils
-from torchvision import transforms
-def stack_grayscale_tensor(tensor):
- tensor = torch.cat([tensor, tensor, tensor], 0)
- return tensor
+def get_list_from_filenames(file_path):
+ # input: relative path to .txt file with file names
+ # output: list of relative path names
+ with open(file_path) as f:
+ lines = f.read().splitlines()
+ return lines
+
+class Synhead(Dataset):
+ def __init__(self, data_dir, csv_path, transform, test=False):
+ column_names = ['path', 'bbox_x_min', 'bbox_y_min', 'bbox_x_max', 'bbox_y_max', 'yaw', 'pitch', 'roll']
+ tmp_df = pd.read_csv(csv_path, sep=',', names=column_names, index_col=False, encoding="utf-8-sig")
+ self.data_dir = data_dir
+ self.transform = transform
+ self.X_train = tmp_df['path']
+ self.y_train = tmp_df[['bbox_x_min', 'bbox_y_min', 'bbox_x_max', 'bbox_y_max', 'yaw', 'pitch', 'roll']]
+ self.length = len(tmp_df)
+ self.test = test
+
+ def __getitem__(self, index):
+ path = os.path.join(self.data_dir, self.X_train.iloc[index]).strip('.jpg') + '.png'
+ img = Image.open(path)
+ img = img.convert('RGB')
+
+ x_min, y_min, x_max, y_max, yaw, pitch, roll = self.y_train.iloc[index]
+ x_min = float(x_min); x_max = float(x_max)
+ y_min = float(y_min); y_max = float(y_max)
+ yaw = -float(yaw); pitch = float(pitch); roll = float(roll)
+
+ # k = 0.2 to 0.40
+ k = np.random.random_sample() * 0.2 + 0.2
+ x_min -= 0.6 * k * abs(x_max - x_min)
+ y_min -= 2 * k * abs(y_max - y_min)
+ x_max += 0.6 * k * abs(x_max - x_min)
+ y_max += 0.6 * k * abs(y_max - y_min)
+
+ width, height = img.size
+ # Crop the face
+ img = img.crop((int(x_min), int(y_min), int(x_max), int(y_max)))
+
+ # Flip?
+ rnd = np.random.random_sample()
+ if rnd < 0.5:
+ yaw = -yaw
+ roll = -roll
+ img = img.transpose(Image.FLIP_LEFT_RIGHT)
+
+ # Blur?
+ rnd = np.random.random_sample()
+ if rnd < 0.05:
+ img = img.filter(ImageFilter.BLUR)
+
+ # Bin values
+ bins = np.array(range(-99, 102, 3))
+ binned_pose = np.digitize([yaw, pitch, roll], bins) - 1
+
+ labels = torch.LongTensor(binned_pose)
+ cont_labels = torch.FloatTensor([yaw, pitch, roll])
+
+ if self.transform is not None:
+ img = self.transform(img)
+
+ return img, labels, cont_labels, self.X_train[index]
+
+ def __len__(self):
+ return self.length
class Pose_300W_LP(Dataset):
+ # Head pose from 300W-LP dataset
def __init__(self, data_dir, filename_path, transform, img_ext='.jpg', annot_ext='.mat', image_mode='RGB'):
self.data_dir = data_dir
self.transform = transform
@@ -30,16 +96,14 @@
img = Image.open(os.path.join(self.data_dir, self.X_train[index] + self.img_ext))
img = img.convert(self.image_mode)
mat_path = os.path.join(self.data_dir, self.y_train[index] + self.annot_ext)
- shape_path = os.path.join(self.data_dir, self.y_train[index] + '_shape.npy')
- # Crop the face
+ # Crop the face loosely
pt2d = utils.get_pt2d_from_mat(mat_path)
x_min = min(pt2d[0,:])
y_min = min(pt2d[1,:])
x_max = max(pt2d[0,:])
y_max = max(pt2d[1,:])
- # k = 0.35 was being used beforehand
# k = 0.2 to 0.40
k = np.random.random_sample() * 0.2 + 0.2
x_min -= 0.6 * k * abs(x_max - x_min)
@@ -71,10 +135,8 @@
bins = np.array(range(-99, 102, 3))
binned_pose = np.digitize([yaw, pitch, roll], bins) - 1
- # Get shape
- shape = np.load(shape_path)
-
- labels = torch.LongTensor(np.concatenate((binned_pose, shape), axis = 0))
+ # Get target tensors
+ labels = binned_pose
cont_labels = torch.FloatTensor([yaw, pitch, roll])
if self.transform is not None:
@@ -87,6 +149,7 @@
return self.length
class Pose_300W_LP_random_ds(Dataset):
+ # 300W-LP dataset with random downsampling
def __init__(self, data_dir, filename_path, transform, img_ext='.jpg', annot_ext='.mat', image_mode='RGB'):
self.data_dir = data_dir
self.transform = transform
@@ -104,9 +167,8 @@
img = Image.open(os.path.join(self.data_dir, self.X_train[index] + self.img_ext))
img = img.convert(self.image_mode)
mat_path = os.path.join(self.data_dir, self.y_train[index] + self.annot_ext)
- shape_path = os.path.join(self.data_dir, self.y_train[index] + '_shape.npy')
- # Crop the face
+ # Crop the face loosely
pt2d = utils.get_pt2d_from_mat(mat_path)
x_min = min(pt2d[0,:])
y_min = min(pt2d[1,:])
@@ -123,12 +185,11 @@
# We get the pose in radians
pose = utils.get_ypr_from_mat(mat_path)
- # And convert to degrees.
pitch = pose[0] * 180 / np.pi
yaw = pose[1] * 180 / np.pi
roll = pose[2] * 180 / np.pi
- ds = np.random.randint(1,11)
+ ds = 1 + np.random.randint(0,4) * 5
original_size = img.size
img = img.resize((img.size[0] / ds, img.size[1] / ds), resample=Image.NEAREST)
img = img.resize((original_size[0], original_size[1]), resample=Image.NEAREST)
@@ -149,98 +210,14 @@
bins = np.array(range(-99, 102, 3))
binned_pose = np.digitize([yaw, pitch, roll], bins) - 1
- # Get shape
- shape = np.load(shape_path)
-
- labels = torch.LongTensor(np.concatenate((binned_pose, shape), axis = 0))
+ # Get target tensors
+ labels = binned_pose
cont_labels = torch.FloatTensor([yaw, pitch, roll])
if self.transform is not None:
img = self.transform(img)
return img, labels, cont_labels, self.X_train[index]
-
- def __len__(self):
- # 122,450
- return self.length
-
-class Pose_300W_LP_SR(Dataset):
- def __init__(self, data_dir, filename_path, transform, img_ext='.jpg', annot_ext='.mat', image_mode='RGB'):
- self.data_dir = data_dir
- self.transform = transform
- self.img_ext = img_ext
- self.annot_ext = annot_ext
-
- filename_list = get_list_from_filenames(filename_path)
-
- self.X_train = filename_list
- self.y_train = filename_list
- self.image_mode = image_mode
- self.length = len(filename_list)
-
- def __getitem__(self, index):
- img = Image.open(os.path.join(self.data_dir, self.X_train[index] + self.img_ext))
- img = img.convert(self.image_mode)
- mat_path = os.path.join(self.data_dir, self.y_train[index] + self.annot_ext)
-
- # Crop the face
- pt2d = utils.get_pt2d_from_mat(mat_path)
- x_min = min(pt2d[0,:])
- y_min = min(pt2d[1,:])
- x_max = max(pt2d[0,:])
- y_max = max(pt2d[1,:])
-
- # k = 0.2 to 0.40
- k = np.random.random_sample() * 0.2 + 0.2
- x_min -= 0.6 * k * abs(x_max - x_min)
- y_min -= 2 * k * abs(y_max - y_min)
- x_max += 0.6 * k * abs(x_max - x_min)
- y_max += 0.6 * k * abs(y_max - y_min)
- img = img.crop((int(x_min), int(y_min), int(x_max), int(y_max)))
-
- # We get the pose in radians
- pose = utils.get_ypr_from_mat(mat_path)
- # And convert to degrees.
- pitch = pose[0] * 180 / np.pi
- yaw = pose[1] * 180 / np.pi
- roll = pose[2] * 180 / np.pi
-
- rnd = np.random.random_sample()
- if rnd < 0.5:
- ds = 10
- original_size = img.size
- img = img.resize((img.size[0] / ds, img.size[1] / ds), resample=Image.NEAREST)
- img = img.resize((original_size[0], original_size[1]), resample=Image.NEAREST)
-
- # Flip?
- rnd = np.random.random_sample()
- if rnd < 0.5:
- yaw = -yaw
- roll = -roll
- img = img.transpose(Image.FLIP_LEFT_RIGHT)
-
- # Blur?
- rnd = np.random.random_sample()
- if rnd < 0.05:
- img = img.filter(ImageFilter.BLUR)
-
- img_ycc = img.convert('YCbCr')
-
- # Bin values
- bins = np.array(range(-99, 102, 3))
- binned_pose = np.digitize([yaw, pitch, roll], bins) - 1
-
- labels = torch.LongTensor(np.concatenate((binned_pose, shape), axis = 0))
- cont_labels = torch.FloatTensor([yaw, pitch, roll])
-
- # Transforms
- img = transforms.Scale(240)(img)
- img = transforms.RandomCrop(224)(img)
- img_ycc = img.convert('YCbCr')
- img = transforms.ToTensor()
- img_ycc = transforms.ToTensor()
-
- return img, img_ycc, labels, cont_labels, self.X_train[index]
def __len__(self):
# 122,450
@@ -265,7 +242,7 @@
img = img.convert(self.image_mode)
mat_path = os.path.join(self.data_dir, self.y_train[index] + self.annot_ext)
- # Crop the face
+ # Crop the face loosely
pt2d = utils.get_pt2d_from_mat(mat_path)
x_min = min(pt2d[0,:])
@@ -301,6 +278,7 @@
return self.length
class AFLW2000_ds(Dataset):
+ # AFLW2000 dataset with fixed downsampling
def __init__(self, data_dir, filename_path, transform, img_ext='.jpg', annot_ext='.mat', image_mode='RGB'):
self.data_dir = data_dir
self.transform = transform
@@ -319,7 +297,7 @@
img = img.convert(self.image_mode)
mat_path = os.path.join(self.data_dir, self.y_train[index] + self.annot_ext)
- # Crop the face
+ # Crop the face loosely
pt2d = utils.get_pt2d_from_mat(mat_path)
x_min = min(pt2d[0,:])
y_min = min(pt2d[1,:])
@@ -333,7 +311,7 @@
y_max += 0.6 * k * abs(y_max - y_min)
img = img.crop((int(x_min), int(y_min), int(x_max), int(y_max)))
- ds = 8
+ ds = 3 # downsampling factor
original_size = img.size
img = img.resize((img.size[0] / ds, img.size[1] / ds), resample=Image.NEAREST)
img = img.resize((original_size[0], original_size[1]), resample=Image.NEAREST)
@@ -358,67 +336,8 @@
# 2,000
return self.length
-class AFLW2000_random_ds(Dataset):
- def __init__(self, data_dir, filename_path, transform, img_ext='.jpg', annot_ext='.mat', image_mode='RGB'):
- self.data_dir = data_dir
- self.transform = transform
- self.img_ext = img_ext
- self.annot_ext = annot_ext
-
- filename_list = get_list_from_filenames(filename_path)
-
- self.X_train = filename_list
- self.y_train = filename_list
- self.image_mode = image_mode
- self.length = len(filename_list)
-
- def __getitem__(self, index):
- img = Image.open(os.path.join(self.data_dir, self.X_train[index] + self.img_ext))
- img = img.convert(self.image_mode)
- mat_path = os.path.join(self.data_dir, self.y_train[index] + self.annot_ext)
-
- # Crop the face
- pt2d = utils.get_pt2d_from_mat(mat_path)
- x_min = min(pt2d[0,:])
- y_min = min(pt2d[1,:])
- x_max = max(pt2d[0,:])
- y_max = max(pt2d[1,:])
-
- k = 0.20
- x_min -= 2 * k * abs(x_max - x_min)
- y_min -= 2 * k * abs(y_max - y_min)
- x_max += 2 * k * abs(x_max - x_min)
- y_max += 0.6 * k * abs(y_max - y_min)
- img = img.crop((int(x_min), int(y_min), int(x_max), int(y_max)))
-
- rnd = np.random.random_sample()
- if rnd < 0.5:
- ds = 10
- original_size = img.size
- img = img.resize((img.size[0] / ds, img.size[1] / ds), resample=Image.NEAREST)
- img = img.resize((original_size[0], original_size[1]), resample=Image.NEAREST)
-
- # We get the pose in radians
- pose = utils.get_ypr_from_mat(mat_path)
- # And convert to degrees.
- pitch = pose[0] * 180 / np.pi
- yaw = pose[1] * 180 / np.pi
- roll = pose[2] * 180 / np.pi
- # Bin values
- bins = np.array(range(-99, 102, 3))
- labels = torch.LongTensor(np.digitize([yaw, pitch, roll], bins) - 1)
- cont_labels = torch.FloatTensor([yaw, pitch, roll])
-
- if self.transform is not None:
- img = self.transform(img)
-
- return img, labels, cont_labels, self.X_train[index]
-
- def __len__(self):
- # 2,000
- return self.length
-
class AFLW_aug(Dataset):
+ # AFLW dataset with flipping
def __init__(self, data_dir, filename_path, transform, img_ext='.jpg', annot_ext='.txt', image_mode='RGB'):
self.data_dir = data_dir
self.transform = transform
@@ -445,7 +364,7 @@
yaw = pose[0] * 180 / np.pi
pitch = pose[1] * 180 / np.pi
roll = pose[2] * 180 / np.pi
- # Something weird with the roll in AFLW
+ # Fix the roll in AFLW
roll *= -1
# Augment
@@ -455,21 +374,6 @@
yaw = -yaw
roll = -roll
img = img.transpose(Image.FLIP_LEFT_RIGHT)
-
- # Blur?
- # rnd = np.random.random_sample()
- # if rnd < 0.05:
- # img = img.filter(ImageFilter.BLUR)
- # if rnd < 0.025:
- # img = img.filter(ImageFilter.BLUR)
- #
- # rnd = np.random.random_sample()
- # if rnd < 0.05:
- # nb = np.random.randint(1,5)
- # img = img.rotate(-nb)
- # elif rnd > 0.95:
- # nb = np.random.randint(1,5)
- # img = img.rotate(nb)
# Bin values
bins = np.array(range(-99, 102, 3))
@@ -513,7 +417,7 @@
yaw = pose[0] * 180 / np.pi
pitch = pose[1] * 180 / np.pi
roll = pose[2] * 180 / np.pi
- # Something weird with the roll in AFLW
+ # Fix the roll in AFLW
roll *= -1
# Bin values
bins = np.array(range(-99, 102, 3))
@@ -557,7 +461,7 @@
line = annot.readline().split(' ')
yaw, pitch, roll = [float(line[1]), float(line[2]), float(line[3])]
- # Crop the face
+ # Crop the face loosely
k = 0.32
x1 = float(line[4])
y1 = float(line[5])
@@ -633,9 +537,11 @@
R = R[:3,:]
pose_annot.close()
- roll = np.arctan2(R[1][0], R[0][0]) * 180 / np.pi
- yaw = np.arctan2(-R[2][0], np.sqrt(R[2][1] ** 2 + R[2][2] ** 2)) * 180 / np.pi
- pitch = -np.arctan2(R[2][1], R[2][2]) * 180 / np.pi
+ R = np.transpose(R)
+
+ roll = -np.arctan2(R[1][0], R[0][0]) * 180 / np.pi
+ yaw = -np.arctan2(-R[2][0], np.sqrt(R[2][1] ** 2 + R[2][2] ** 2)) * 180 / np.pi
+ pitch = np.arctan2(R[2][1], R[2][2]) * 180 / np.pi
# Loosely crop face
k = 0.35
@@ -660,11 +566,3 @@
def __len__(self):
# 15,667
return self.length
-
-
-def get_list_from_filenames(file_path):
- # input: relative path to .txt file with file names
- # output: list of relative path names
- with open(file_path) as f:
- lines = f.read().splitlines()
- return lines
--
Gitblit v1.8.0