Faster way to patchify a picture to overlapping blocks

2018-11-08 06:33 Moran Reznik imported from Stackoverflow

I'm looking for a FAST (and if possible memory afficiant) way to rewrite a function I crerated as part of Visual bag of words algorithm:

def get_pic_patches(pic, l, s):  # "s" stands for stride

    r, c = pic.shape
    i, j = [0, 0]
    x_range = list(range(0, r, s  ) )
    y_range = list(range(0, c ,  s ) )
    patches = []
    patches_location = []
    for x in x_range:  # without last two since it will exceed dimensions
        for y in y_range:  # without last two since it will exceed dimensions
            if x+ l<= r and y+l <= c:
                patch = pic[x:x +  l , y:y + l ]
                patches_location.append([x, y])  # patch location is the upper left pixel
                patches.append(  patch   )

    return patches, patches_location

it takes a grayscale image (NOT RGB!), desired patch length and stride value, and gives back all patches as a list of numpy array.

On other qestions, I found this:

def patchify(img, patch_shape):
    img = np.ascontiguousarray(img)  # won't make a copy if not needed
    X, Y = img.shape
    x, y = patch_shape
    shape = ((X-x+1), (Y-y+1), x, y) # number of patches, patch_shape 
    strides = img.itemsize*np.array([Y, 1, Y, 1])
    return np.lib.stride_tricks.as_strided(img, shape=shape, strides=strides)

in order to get to return a list, I used it like this:

def patchify(img, patch_shape):
    img = np.ascontiguousarray(img)  # won't make a copy if not needed
    X, Y = img.shape
    x, y = patch_shape
    shape = ((X-x+1), (Y-y+1), x, y) # number of patches, patch_shape 
    strides = img.itemsize*np.array([Y, 1, Y, 1])
    patches = np.lib.stride_tricks.as_strided(img, shape=shape, strides=strides)
    a,b,c,d = patches.shape
    patches = patches.reshape(((a*b),c,d))
    patches = patches.tolist()
    return

but this was actually much slower than my original function! another problem is that is only works with stride = 1, and I want to be able to use all sorts of stride values.

See also questions close to this topic

Replace and Iterate over x- and y-axis of multi-dimensional np.matrix
I have written the following script:
```
import numpy as np

d = 0.05  
limits = np.arange(0, 0.5, d)
matrix = np.zeros([limits.shape[0], limits.shape[0]])
matrix[:, 0] = limits
matrix[0, :] = limits
```
This gives me the following matrix:

How can I make sure that the values in the first column and first row are the "index"-values? So instead of [0,1,2,...,9] there should be [0, 0.05, ...0.45]

Thank you

Concatenate arrays in Python - output error

I am writing a program to concatenate two numpy arrays and I want the program to print a message for each possible outcome (horizontal, vertical, or no concatenation). I have the following code and I don´t understand why even when the first condition (np.hstack) is met it continues evaluating the subsequent if and else statements and wrongly prints that there is both a horizontal concatenation (first condition is met) and that a concatenation is not possible.

import numpy as np
def fun1(a,b):
    if a.shape[0] == b.shape[0]:
        print("The horizontal concatenation is:", np.hstack((a,b)))
    if a.shape[1] == b.shape[1]:
        print("The vertical concatenation is:", np.vstack((a,b)))
    else:
        print("These arrays cannot be concatenated.")

a = np.floor(10*np.random.random((3,2)))
b = np.floor(10*np.random.random((3,4)))
fun1(a,b)

Output:

The horizontal concatenation is: [[5. 0. 1. 1. 3. 7.]
                                  [4. 1. 8. 3. 1. 9.]
                                  [9. 1. 5. 7. 0. 0.]]
These arrays cannot be concatenated.

How to merge different time-series plot into a 2D in Python

I have three different time-series data of the following format where the first column is timestamp and the second column is the value.

0.086206438,10
0.086425551,12
0.089227066,20
0.089262508,24
0.089744425,30
0.090036815,40
0.090054172,28
0.090377569,28
0.090514071,28
0.090762872,28
0.090912691,27

For reproduciability, I have shared the three time-series data I am using here.

From column 2, I wanted to read current row and compare it with the value of the previous row. If it is greater, I keep comparing. If the current value is smaller than the previous row's value continue, I take the difference. Let me make it clear. For example in the above sample record I provided, the seventh row (28) is smaller than the value in the sixth row (40) - so it will be (40-28 =12).

Here is my sample code.

import numpy as np
import pandas as pd
import csv
import numpy as np
import scipy.stats
import matplotlib.pyplot as plt
import seaborn as sns
from scipy.stats import norm
from statsmodels.graphics.tsaplots import plot_acf, acf


protocols = {}


types = {"data1": "data1.csv", "data2": "data2.csv", "data3": "data3.csv"}

for protname, fname in types.items():
    arr = []  
    arr1 = []  
    with open(fname, mode='r', encoding='utf-8-sig') as f:
        reader = csv.reader(f, delimiter=",")
        for i in reader:
            arr.append(int(i[1])) 
            arr1.append(float(i[0])) 

    arr, arr1 = np.array(arr), np.array(arr1)
    diffs = np.diff(arr)
    diffs1 = np.diff(arr1)
    diffs1 = diffs1[diffs > 0] 
    diffs = diffs[diffs > 0] # To keep only the increased values
    protocols[protname] = {
        "rtime": np.array(arr1),
        "rdata": np.array(arr),
        "data": diffs,
        "timediff": diffs,
    }

## change in time 
for protname, values in protocols.items():
    d = values["rdata"]
    t = values["rtime"]
    d = np.diff(d, 1) #/ d[:-1]
    t = np.diff(t, 1)
    plt.plot(t, d, ".", label=protname, alpha=0.5)
    plt.xlabel("Changes in time")
    plt.ylabel("differences")
    plt.legend()
    plt.show()

This gives me the following plots

How can we plot the differences versus the change in time (column onw) in a two-dimensitional (2D) graph of the three data I included separately?

Word2vec compact models

Tell me if there are any w2v models that do not require a dictionary. So, everything that I found in torchtext first wants to know the dictionary build_vocab. But if I have a huge body of text, I would like to have a model that works at the level of phrases. But I did not find one.
supervised learning for parcours

for my school project i got to implement a neural network for a parcours. I know it useless but i want the neural net to learn a simple algorithm:

if front right is bigger than front left -> go right, else -> go left.

I wanna use supervised learning. I got 2 inputs neurons, 2 hidden neurons and 1 output neuron. The goal is that when the player has to go left the output gives a number under 0.5 and if the player has to go right the nn has to return a number greater that 0.5.

Somehow I made a mistake and the nn always tries to return 0.5. Do so know what i did wrong and what i can do now.

thats how the parcours looks like
Categorical Variables and too many NA for ML model

We have a data set of 250 variables and 50,000 records. One variable is numeric, 248 variables are categorical and one variable is binary (the target variable). Each category variable has more than 3000 levels. We have many NA. Each row is the record of diseases that a patient has suffered. That's why there are so many NAs. Because a patient may have suffered 100 diseases, and another has suffered only one. The objective is to be able to predict if patients can have a specific disease from the information of other diseases they have suffered. How can this data set be handled in machine learning?
How to remove high frequency regions from a canny image?

Assume I have a canny detected image, for example this one I stole from the internet:

I need to detect and remove regions with high frequency features. So this image should become soemthing like:

(Excuse the GIMP skills)

In other words, areas containing high variability of of features should just go to black.

To my understanding I could use something like fourier transforms to filter out these kinds of regions, but I am not sure how or if this is the best approach.
Per image normalization vs Overall dataset normalization

I have a datasets of 1000 image.Using cnn For fingure gesture recognition. Should I normalize the Image by finding mean of that image only or the mean of entire dataset...and also suggest which library to use in python for the same
Xcode 10.0 OpenCV Cpp project does not read NSCameraDescription from info.plist

I have already added a info.plist file, but it just wont let me use the webcam.I am in a fix as I have already tried editing the info.plist as a source file. Correct me if I am missing something. Also linked the info.plist to the plist for the app in general settings. Am I missing something? Link the the error is given in the blue highlighted section

Related questions

Faster way to patchify a picture to overlapping blocks

See also questions close to this topic