from tsne import bh_sne
import numpy as np
from skimage.transform import resize
from sklearn.manifold import TSNE
from matplotlib import pyplot as plt
from matplotlib.offsetbox import AnnotationBbox, OffsetImage
import matplotlib.image as mpimg
import warnings
warnings.filterwarnings("ignore")
def gray_to_color(img):
if len(img.shape) == 2:
img = np.dstack((img, img, img))
return img
def min_resize(img, size):
"""
Resize an image so that it is size along the minimum spatial dimension.
"""
w, h = map(float, img.shape[:2])
if min([w, h]) != size:
if w <= h:
img = resize(img, (int(round((h/w)*size)), int(size)))
else:
img = resize(img, (int(size), int(round((w/h)*size))))
return img
def fit(features):
features = np.copy(features).astype('float64')
model = TSNE(n_components=2, n_iter=1000, perplexity=10.0)
f2d = model.fit_transform(features)
return f2d
def image_scatter(features, images, img_res, filenames, res=15000, cval=1.):
"""
Embeds images via tsne into a scatter plot.
Parameters
---------
features: numpy array
Features to visualize
images: list or numpy array
Corresponding images to features. Expects float images from (0,1).
img_res: float or int
Resolution to embed images at
res: float or int
Size of embedding image in pixels
cval: float or numpy array
Background color value
Returns
------
canvas: numpy array
Image of visualization
"""
f2d = fit(features)
images = [gray_to_color(image) for image in images]
images = [min_resize(image, img_res) for image in images]
max_width = max([image.shape[0] for image in images])
max_height = max([image.shape[1] for image in images])
xx = f2d[:, 0]
yy = f2d[:, 1]
x_min, x_max = xx.min(), xx.max()
y_min, y_max = yy.min(), yy.max()
# Fix the ratios
sx = (x_max-x_min)
sy = (y_max-y_min)
if sx > sy:
res_x = sx/float(sy)*res
res_y = res
else:
res_x = res
res_y = sy/float(sx)*res
canvas = np.ones((res_x+max_width, res_y+max_height, 3))*cval
x_coords = np.linspace(x_min, x_max, res_x)
y_coords = np.linspace(y_min, y_max, res_y)
for x, y, image, name in zip(xx, yy, images, filenames):
w, h = image.shape[:2]
x_idx = np.argmin((x - x_coords)**2)
y_idx = np.argmin((y - y_coords)**2)
if name.endswith("8238.jpg"):
print("%s is at (%d, %d)" % (name, y_idx, x_idx))
canvas[x_idx:x_idx+w, y_idx:y_idx+h] = image
plt.figure(figsize=(10,8))
plt.imshow(canvas)
plt.show()
raw_input()
def ab_plotter(xcoords, ycoords, images, labels):
ax = plt.subplot(111)
ax.set_xlim([-30, 30])
ax.set_ylim([-30, 30])
for x, y, i, l in zip(xcoords, ycoords, images, labels):
arr_hand = i
imagebox = OffsetImage(arr_hand, zoom=.1)
xy = [x, y] # coordinates to position this image
ab = AnnotationBbox(imagebox, xy,
xybox=(10., -10.),
xycoords='data',
boxcoords="offset points",
pad=0.0)
ax.annotate(ab, xy = xy)
# rest is just standard matplotlib boilerplate
ax.grid(True)
plt.show()
if __name__ == "__main__":
x_data = np.loadtxt("eng256-initial_hidden_features", delimiter=",")
#x_data = np.loadtxt("eng256-ger256-multilingual_initial_hidden_features", delimiter=",")
y_data = np.loadtxt("val_images", dtype=str)
y_data = [x.replace("\n","") for x in y_data]
images = []
for y in y_data:
images.append(mpimg.imread(y))
canvas_mono = image_scatter(x_data, images, 200, y_data, res=15000)
#canvas_multi = image_scatter(x_data2, images, 100, y_data)
#coords = fit(x_data)
#ab_plotter(coords[:,0], coords[:,1], images, y_data)
