Related
I've to plot multiple scatter and table in a grid space and I'm having a couple of issues with the relative position but most important with defining and maintaining the aspect ratio of the scatter plot.
I've written a script with "fake" data on it to describe my problem and a minimum "not working" example below.
What I have is a dataframe with x, and y positions of objects, and what I want to do is to put the corresponding image below.
Since the image can have an arbitrary aspect ratio I need to read the aspect ratio and construct the scatter plot in that way but I'm unable to make it work.
Another problem is connected with the invert_xaxis and invert_yaxis that don't work (I need that command since the scatter data are inverted.
I've used the following commands, and as far as I've understood each of them should block the aspect ratio of the scatter plot to the same ratio of the figure but they do not work.
The aspect ratio becomes corrected only when the figure is plotted but that eliminates the effect of axis inversion.
I've had a similar problem with setting the aspect ratio of plots without the addition of a figure, sometimes it worked but not with tight_layout.
It is obvious that I'm missing something important....but I'm unable to figure it out.
This is the fake data code:
###############################################################################
# fake data
#general data aspect ratio
image_height= 5 #4270
image_width = 10 # 8192
pix2scale = 0.3125
data_AR = image_height / image_width
#random data generation
data_width = image_width* pix2scale
data_height = image_height * pix2scale
data1x = np.random.uniform(-data_width/2, data_width/2, size=(40))
data1y = np.random.uniform(-data_height/2, data_height/2, size=(40))
data2x = np.random.uniform(-data_width/2, data_width/2, size=(40))
data2y = np.random.uniform(-data_height/2,data_height/2, size=(40))
temp_df1 = pd.DataFrame([data1x,data1y,['random1']*40],index = ['x','y','label']).T
temp_df2 = pd.DataFrame([data2x,data2y,['random2']*40],index = ['x','y','label']).T
df = pd.concat([temp_df1,temp_df2],axis = 0, ignore_index = True)
del temp_df1, temp_df2
#sample image generation of variable aspect ratio
img_size = (image_height, image_width)
R_layer = np.ones(shape= img_size)*0.50
G_layer = np.ones(shape= img_size)*0.50
B_layer = np.ones(shape= img_size)*0.50
A_layer = np.ones(shape= img_size)
img = np.dstack([R_layer,G_layer,B_layer,A_layer])
#add a mark at the top left of the image
for k in range(0,3):
for i in range(0,int(image_width*0.2*data_AR)):
for j in range(0,int(image_width*0.2)):
img[i,j,k] = 0
#add a mark at the center of the image
# get center coordinates of the image
center = [[2, 4], [2, 5]]
for k in range(0,3):
for point in center:
if k == 0:
img[point[0],point[1],k] = 1
else:
img[point[0],point[1],k] = 0
#show image
fig, ax = plt.subplots()
ax.imshow(img)
###############################################################################
this is the code that generates the image:
#%%
# sample code
# at this point Iìve already loaded the image, the pix2scale value
# and the df containing data points
#read image aspect ratio
img_AR = img.shape[0]/img.shape[1]
pixel_width = img.shape[1]
pixel_height = img.shape[0]
# each pixel correspond to 0.3125 unit (mm)
pix2scale = 0.3125
#define image position
#the center of the image need to be placed at (0,0)
#bottom left corner
left = - (pixel_width * pix2scale)/2
bottom = - (pixel_height * pix2scale)/2
right = left + (pixel_width * pix2scale)
top = bottom + (pixel_height * pix2scale)
extent = [left,right,bottom,top]
#figure definition
figure_width = 15 #inch
figure_AR = 1
scatter_AR = img_AR
#initialize figure
fig_s= plt.figure(figsize = (figure_width,figure_width*figure_AR))
gs = plt.GridSpec (3,3)
#scatter plot in ax1
ax1 = fig_s.add_subplot(gs[:2,:2])
g = sns.scatterplot( data = df,
x = 'x',
y = 'y',
hue = 'label',
ax =ax1
)
ax1.invert_xaxis()
ax1.invert_yaxis()
#resize the figure box
box = ax1.get_position()
ax1.set_position([box.x0,box.y0,box.width*0.4,box.width*0.4*scatter_AR])
ax1.legend(loc = 'center left', bbox_to_anchor = (1,0.5))
ax1.set_title('Inclusions Scatter Plot')
ax1.set_aspect(scatter_AR)
#plt image
ax1.imshow(img,extent = extent)
#scatter plot
ax2 = fig_s.add_subplot(gs[2,:2])
g = sns.scatterplot( data = df,
x = 'x',
y = 'y',
hue = 'label',
ax =ax2
)
#resize the figure box
box = ax2.get_position()
ax2.set_position([box.x0,box.y0,box.width*0.4,box.width*0.4*scatter_AR])
ax2.legend(loc = 'center left', bbox_to_anchor = (1,0.5))
ax2.set_title('Inclusions Scatter Plot')
ax2.set_aspect(scatter_AR)
ax2.imshow(img,extent = extent)
#scatter plot
ax3 = fig_s.add_subplot(gs[1,2])
g = sns.scatterplot( data = df,
x = 'x',
y = 'y',
hue = 'label',
ax =ax3
)
#resize the figure box
box = ax3.get_position()
ax3.set_position([box.x0,box.y0,box.width*0.4,box.width*0.4*scatter_AR])
ax3.legend(loc = 'center left', bbox_to_anchor = (1,0.5))
ax3.set_title('Inclusions Scatter Plot')
ax3.set_aspect(scatter_AR)
ax3.imshow(img,extent = extent)
#add suptitle to figure
fig_s.suptitle('my title',fontsize= 22)
fig_s.subplots_adjust(top=0.85)
# #make it fancy
for i in range(3):
fig_s.tight_layout()
plt.pause(0.2)
I've plotted multiple grid because I wanted to test the tight_layout().
[enter image description here][2]
I was working on this https://www.kaggle.com/gti-upm/leapgestrecog data set lately. Its a hand gesture dataset and I was trying to make a classifier. Due to images available in different types of folder I made my on data loader. Here it is
class DatasetLoader(Dataset):
def __init__(self,path):
self.path_list = path
self.labels = []
self.to_tensor = transforms.ToTensor()
self.resize = transforms.Resize((120,320))
self.gray = transforms.Grayscale(num_output_channels = 1)
self._init_dataset()
def _init_dataset(self):
labels = set()
for diro in os.listdir("/kaggle/input/leapgestrecog/leapGestRecog"):
for d in os.listdir(os.path.join("/kaggle/input/leapgestrecog/leapGestRecog",diro)):
if len(d.split('_'))>2:
labels.add("_".join(d.split("_")[-2:]))
else:
labels.add(d.split("_")[-1])
labels = list(labels)
## help me on this line with some codes
def __getitem__(self,idx):
if torch.is_tensor(idx):
idx = idx.tolist()
img_name = self.path_list[idx]
img = Image.open(img_name)
img = self.resize(img)
img = self.gray(img)
img = self.to_tensor(img)
if len(img_name.split('/')[-2].split('_')) > 2:
label = "_".join(img_name.split('/')[-2].split('_')[-2:])
else:
label = img_name.split('/')[-2].split('_')[-1]
label = ## Here also
return img,label
def __len__(self):
return len(self.path_list)
I have problem with label which I am getting from this dataset loader. As I have created a model which takes n batches of data with 10 classes so during loss calculation I need my labels to of size(n,10). I dont know what to do. Here is my network design:
class Net(nn.Module):
def __init__(self):
super(Net,self).__init__()
self.conv1 = nn.Conv2d(1,32,5)
self.pool = nn.MaxPool2d(2,2)
self.conv2 = nn.Conv2d(32,64,3)
self.conv3 = nn.Conv2d(64,64,3)
self.fc1 = nn.Linear(64*38*13,128)
self.fc2 = nn.Linear(128,10)
def forward(self,x):
x = self.pool(F.relu(self.conv1(x)))
x = self.pool(F.relu(self.conv2(x)))
x = self.pool(F.relu(self.conv3(x)))
x = x.view(64,64*38*13)
x = F.relu(self.fc1(x))
return F.log_softmax(self.fc2(x),dim = 1)
If y is label of an image. To train our network we feed loss function with y and output. But output we get is of size (64,10) , so I need help with the label in dataloader
I see that you have a bit of misunderstanding about the input dimensions to a multi-class loss function in PyTorch. The most commonly used loss function for classification problems is nn.CrossEntropyLoss(), which expects raw logits of size (n, c) (e.g. (64, 10)) as input1, and target (ground truth label) of size (n) (e.g. (10)).
So instead of doing return F.log_softmax(self.fc2(x),dim = 1), it would be more stable to directly do return x and use CrossEntropyLoss. There is no need to reshape your labels, you can directly calculate the loss by doing something like this:
criterion = nn.CrossEntropyLoss()
# Let x be (64, 10) output from model
# Let y be (10,) label
loss = criterion(x, y)
I'm currently working on traffic jams analysis and was wondering if there's a way to animate the generation of a plot of such jams.
A plot of this things grow from up to the lower end of the figure, each 'row' is a time instance. The horizontal axis is just the road indicating at each point the position of each vehicle and, with a certain numeric value, the velocity of it. So applying different colors to different velocities, you get a plot that shows how a jam evolves through time in a given road.
My question is, how can I use matplotlib to generate an animation of each instance of the road in time to get such a plot?
The plot is something like this:
I'm simulating a road with vehicles with certain velocities through time, so I wish to animate a plot showing how the traffic jams evolve...
EDIT:
I add some code to make clear what I'm already doing
import numpy as np
from matplotlib import pyplot as plt
from matplotlib import animation, rc
plt.rcParams['animation.ffmpeg_path'] = u'/usr/bin/ffmpeg'
# model params
vmax = 5
lenroad = 50
prob = 0.4
# sim params
numiters = 10
# traffic model
def nasch():
gaps = np.full(road.shape, -1)
road_r4 = np.full(road.shape, -1)
for n,x in enumerate(road):
if x > -1:
d = 1
while road[(n+d) % len(road)] < 0:
d += 1
d -= 1
gaps[n] = d
road_r1 = np.where(road!=-1, np.minimum(road+1, vmax), -1)
road_r2 = np.where(road_r1!=-1, np.minimum(road_r1, gaps), -1)
road_r3 = np.where(road_r2!=-1, np.where(np.random.rand() < prob, np.maximum(road-1, 0), road), -1)
for n,x in enumerate(road_r3):
if x > -1:
road_r4[(n+x) % len(road_r3)] = x
return road_r4
def plot_nasch(*args):
road = nasch()
plot.set_array([road])
return plot,
# init road
road = np.random.randint(-10, vmax+1, [lenroad])
road = np.where(road>-1, road, -1)
# simulate
fig = plt.figure()
plot = plt.imshow([road], cmap='Pastel2', interpolation='nearest')
for i in range(numiters):
ani = animation.FuncAnimation(fig, plot_nasch, frames=100, interval=500, blit=True)
plt.show()
And I get the following figure, just one road instead of each road painted at the bottom of the previous one:
This is possibly what you want, although I'm not sure why you want to animate the time, since time is already one of the axes in the plot.
The idea here is to store the simulation results of a time-step row by row in an array and replot this array. Thereby previous simulation results are not lost.
import numpy as np
from matplotlib import pyplot as plt
from matplotlib import animation, rc
# model params
vmax = 5
lenroad = 50
prob = 0.4
# sim params
numiters = 25
# traffic model
def nasch():
global road
gaps = np.full(road.shape, -1)
road_r4 = np.full(road.shape, -1)
for n,x in enumerate(road):
if x > -1:
d = 1
while road[(n+d) % len(road)] < 0:
d += 1
d -= 1
gaps[n] = d
road_r1 = np.where(road!=-1, np.minimum(road+1, vmax), -1)
road_r2 = np.where(road_r1!=-1, np.minimum(road_r1, gaps), -1)
road_r3 = np.where(road_r2!=-1, np.where(np.random.rand() < prob, np.maximum(road-1, 0), road), -1)
for n,x in enumerate(road_r3):
if x > -1:
road_r4[(n+x) % len(road_r3)] = x
return road_r4
def plot_nasch(i):
print i
global road
road = nasch()
#store result in array
road_over_time[i+1,:] = road
# plot complete array
plot.set_array(road_over_time)
# init road
road = np.random.randint(-10, vmax+1, [lenroad])
road = np.where(road>-1, road, -1)
# initiate array
road_over_time = np.zeros((numiters+1, lenroad))*np.nan
road_over_time[0,:] = road
fig = plt.figure()
plot = plt.imshow(road_over_time, cmap='Pastel2', interpolation='nearest', vmin=-1.5, vmax=6.5)
plt.colorbar()
ani = animation.FuncAnimation(fig, plot_nasch, frames=numiters, init_func=lambda : 1, interval=400, blit=False, repeat=False)
plt.show()
I'm trying to code a program that can take text and animate it to bounce on a loop, like a ball bouncing to the floor. I used a similar piece of code I found a starting point as I'm still fairly new to Pygame (thank you Pete Shinners, whoever you are), but after updating the code and playing with it for a long time I still can't get it to blit to the screen correctly. The text starts above the rendered area and then gradually falls into view, but the top part of the text is cut off.
I've tried moving the blitted region around the window and resizing the rectangles and surface the program is using, but nothing seems to fix it.
import os, sys, math, pygame, pygame.font, pygame.image
from pygame.locals import *
def bounce():
# define constants
G = 0.98
FLOOR = 0
COEFFICIENT = 0.8
#define variables
ball = 500
direction = 'DOWN'
v = 0
count = 0
#create array to store data
array = [ball]
while True:
if count == 4:
return array
elif ball > FLOOR and direction == 'DOWN':
v += G
if (ball - v) >= FLOOR:
ball = ball - v
array.append(round(ball,2))
else:
ball = FLOOR
array.append(round(ball,2))
direction = 'UP'
v *= COEFFICIENT
count += 1
elif ball >= FLOOR and direction == 'UP':
v -= G
if (ball + v) >= FLOOR:
ball = ball + v
array.append(round(ball,2))
if v <= 0:
direction = 'DOWN'
else:
ball = FLOOR
array.append(ball)
direction = 'UP'
v *= COEFFICIENT
class textBouncy:
array = bounce()
def __init__(self, font, message, fontcolor, amount=10):
# Render the font message
self.base = font.render(message, 0, fontcolor)
# bounce amount (height)
self.amount = amount
#size = rect of maximum height/width of text
self.size = self.base.get_rect().inflate(0, amount).size
#normalise array to meet height restriction
self.array = [round(-x/(500/amount),2) for x in array]
def animate(self):
# create window surface s
s = pygame.Surface(self.size)
# height = max inflated height
height = self.size[1]
# define a step-sized rectangle in the location of the step
src = Rect(0, 0, self.base.get_width(), height)
# moves the message according to the array list.
dst = src.move(0, self.array[i])
if (i + 1) == len(self.array):
global i
i = 0
# blits the information onto the screen
s.blit(self.base, dst, src)
return s
entry_info = 'Bouncing ball text'
if __name__ == '__main__':
pygame.init()
#create text renderer
i = 0
array = bounce()
bigfont = pygame.font.Font(None, 60)
white = 255, 255, 255
renderer = textBouncy(bigfont, entry_info, white, 16)
text = renderer.animate()
#create a window the correct size
win = pygame.display.set_mode(text.get_size())
win.blit(text, (0, 10))
pygame.display.flip()
#run animation loop
finished = 0
while True:
pygame.time.delay(10)
for event in pygame.event.get():
if event.type == QUIT:
pygame.quit()
sys.exit()
text = renderer.animate()
i += 1
win.blit(text, (0, 10)) # blits the finished product from animate
pygame.display.flip()
(Quote) "it all comes down to math really" Kay so you need to - the y axis when you want to make it go up and + the x axis to make it go side ways you could make it go up and down will moveing it horizontally and then when it reaches a point it will stop moving horizontally and just bonce up and down +ing it more every time
That was my 100$ which took me 5 mins to write
After revisiting this I managed to work this out - I needed to add everything I blitted down to compensate for the bounce up. So in the __init__function:
self.array = [round(-x/(500/amount),2)**+self.amount** for x in array]
Works perfectly now :)
using Hough Transform, how can I detect and get coordinates of (x0,y0) and "a" and "b" of an ellipse in 2D space?
This is ellipse01.bmp:
I = imread('ellipse01.bmp');
[m n] = size(I);
c=0;
for i=1:m
for j=1:n
if I(i,j)==1
c=c+1;
p(c,1)=i;
p(c,2)=j;
end
end
end
Edges=transpose(p);
Size_Ellipse = size(Edges);
B = 1:ceil(Size_Ellipse(1)/2);
Acc = zeros(length(B),1);
a1=0;a2=0;b1=0;b2=0;
Ellipse_Minor=[];Ellipse_Major=[];Ellipse_X0 = [];Ellipse_Y0 = [];
Global_Threshold = ceil(Size_Ellipse(2)/6);%Used for Major Axis Comparison
Local_Threshold = ceil(Size_Ellipse(1)/25);%Used for Minor Axis Comparison
[Y,X]=find(Edges);
Limit=numel(Y);
Thresh = 150;
Para=[];
for Count_01 =1:(Limit-1)
for Count_02 =(Count_01+1):Limit
if ((Count_02>Limit) || (Count_01>Limit))
continue
end
a1=Y(Count_01);b1=X(Count_01);
a2=Y(Count_02);b2=X(Count_02);
Dist_01 = (sqrt((a1-a2)^2+(b1-b2)^2));
if (Dist_01 >Global_Threshold)
Center_X0 = (b1+b2)/2;Center_Y0 = (a1+a2)/2;
Major = Dist_01/2.0;Alpha = atan((a2-a1)/(b2-b1));
if(Alpha == 0)
for Count_03 = 1:Limit
if( (Count_03 ~= Count_01) || (Count_03 ~= Count_02))
a3=Y(Count_03);b3=X(Count_03);
Dist_02 = (sqrt((a3 - Center_Y0)^2+(b3 - Center_X0)^2));
if(Dist_02 > Local_Threshold)
Cos_Tau = ((Major)^2 + (Dist_02)^2 - (a3-a2)^2 - (b3-b2)^2)/(2*Major*Dist_02);
Sin_Tau = 1 - (Cos_Tau)^2;
Minor_Temp = ((Major*Dist_02*Sin_Tau)^2)/(Major^2 - ((Dist_02*Cos_Tau)^2));
if((Minor_Temp>1) && (Minor_Temp<B(end)))
Acc(round(Minor_Temp)) = Acc(round(Minor_Temp))+1;
end
end
end
end
end
Minor = find(Acc == max(Acc(:)));
if(Acc(Minor)>Thresh)
Ellipse_Minor(end+1)=Minor(1);Ellipse_Major(end+1)=Major;
Ellipse_X0(end+1) = Center_X0;Ellipse_Y0(end+1) = Center_Y0;
for Count = 1:numel(X)
Para_X = ((X(Count)-Ellipse_X0(end))^2)/(Ellipse_Major(end)^2);
Para_Y = ((Y(Count)-Ellipse_Y0(end))^2)/(Ellipse_Minor(end)^2);
if (((Para_X + Para_Y)>=-2)&&((Para_X + Para_Y)<=2))
Edges(X(Count),Y(Count))=0;
end
end
end
Acc = zeros(size(Acc));
end
end
end
Although this is an old question, perhaps what I found can help someone.
The main problem of using the normal Hough Transform to detect ellipses is the dimension of the accumulator, since we would need to vote for 5 variables (the equation is explained here):
There is a very nice algorithm where the accumulator can be a simple 1D array, for example, and that runs in . If you wanna see code, you can look at here (the image used to test was that posted above).
If you use circle for rough transform is given as rho = xcos(theta) + ysin(theta)
For ellipse since it is
You could transform the equation as
rho = axcos(theta) + bysin(theta)
Although I am not sure if you use standard Hough Transform, for ellipse-like transforms, you could manipulate the first given function.
If your ellipse is as provided, being a true ellipse and not a noisy sample of points;
the search for the two furthest points gives the ends of the major axis,
the search for the two nearest points gives the ends of the minor axis,
the intersection of these lines (you can check it's a right angle) occurs at the centre.
If you know the 'a' and 'b' of an ellipse then you can rescale the image by factor of a/b in one direction and look for circle. I am still thinking about what to do when a and b are unknown.
If you know that it is circle then use Hough transform for circles. Here is a sample code:
int accomulatorResolution = 1; // for each pixel
int minDistBetweenCircles = 10; // In pixels
int cannyThresh = 20;
int accomulatorThresh = 5*_accT+1;
int minCircleRadius = 0;
int maxCircleRadius = _maxR*10;
cvClearMemStorage(storage);
circles = cvHoughCircles( gryImage, storage,
CV_HOUGH_GRADIENT, accomulatorResolution,
minDistBetweenCircles,
cannyThresh , accomulatorThresh,
minCircleRadius,maxCircleRadius );
// Draw circles
for (int i = 0; i < circles->total; i++){
float* p = (float*)cvGetSeqElem(circles,i);
// Draw center
cvCircle(dstImage, cvPoint(cvRound(p[0]),cvRound(p[1])),
1, CV_RGB(0,255,0), -1, 8, 0 );
// Draw circle
cvCircle(dstImage, cvPoint(cvRound(p[0]),cvRound(p[1])),
cvRound(p[2]),CV_RGB(255,0,0), 1, 8, 0 );
}