Feature extraction and take color histogram

Question

I am working on an image processing feature extraction. I have a photo of a bird in which I have to extract bird area and tell what color the bird has. I used canny feature extraction method to get the edges of a bird.

How to extract only bird area and make the background to blue color?

openCv solution should also be fine.

import skimage
import numpy as np
%matplotlib inline
import matplotlib.pyplot as plt

import os
filename = os.path.join(os.getcwd(),'image\image_bird.jpeg')
from skimage import io
bird =io.imread(filename,as_grey=True)
plt.imshow(bird)

from skimage import feature
edges = feature.canny(bird,sigma=1)
plt.imshow(edges )

Actual bird image can be taken from bird link

I want to extract the bird from the image and then create a histogram of bird color @desertnaut — Sumeet Kumar Yadav, Sep 03 '18 at 17:16
Is the background always blurred or more or less uniformly colored? — SilverMonkey, Sep 03 '18 at 23:37
Background is uniform coloured , but it could be of any color @SilverMonkey — Sumeet Kumar Yadav, Sep 04 '18 at 02:54
Try to segment the image using [normalized-cut](http://scikit-image.org/docs/dev/auto_examples/segmentation/plot_ncut.html#sphx-glr-auto-examples-segmentation-plot-ncut-py), then take the labels that overlap the edge map you found with Canny. This way, you can prepare a foreground mask from which you can extract the pixels you are interested in, to calculate the histogram. If you just want to color the background, you can easily do it using the foreground mask. — dhanushka, Sep 04 '18 at 04:23
I want to fully automate, but whats your idea about semi-automatic ?@SilverMonkey — Sumeet Kumar Yadav, Sep 04 '18 at 12:33
A semi-automatic approach would be, in my understanding, an approach that needs certain user inputs other then the image to function. This could be, for example, (a) starting point(s) and color(s) for background detection etc. Those inputs would make the code much easier. — SilverMonkey, Sep 04 '18 at 12:37
bird color will be mostly black and white and the background will be red . Somthing like https://i.dailymail.co.uk/i/pix/2010/03/28/article-0-08E733E0000005DC-861_634x425.jpg — Sumeet Kumar Yadav, Sep 04 '18 at 12:41
That background is blurry but not uniformly colored as in your first example image. — SilverMonkey, Sep 04 '18 at 12:42
Let us [continue this discussion in chat](https://chat.stackoverflow.com/rooms/179380/discussion-between-sumeet-and-silvermonkey). — Sumeet Kumar Yadav, Sep 04 '18 at 12:43
the background might not be uniform but the bird will be black and white for sure — Sumeet Kumar Yadav, Sep 04 '18 at 12:45
Can I do something like how its done in this example https://stackoverflow.com/questions/40527769/removing-black-background-and-make-transparent-from-grabcut-output-in-python-ope @SilverMonkey — Sumeet Kumar Yadav, Sep 04 '18 at 12:46
After doing feature detection using canny , can I do normalized-cut on that ? Can you give s — Sumeet Kumar Yadav, Sep 04 '18 at 13:07
After doing feature detection using canny , can I do normalized-cut on that ? Can you give some examples @dhanushka — Sumeet Kumar Yadav, Sep 04 '18 at 13:10

SilverMonkey · Accepted Answer · 2018-09-06T16:58:44.777

23

Identify the edges of your image
Binarize the image via automatic thresholding
Use contour detection to identify black regions which are inside a white region and merge them with the white region. (Mockup, image may slightly vary)
Use the created image as mask to color the background and color it This can be done by simply setting each background pixel (black) to its respective color.

As you can see, the approach is far from perfect, but should give you a general idea about how to accomplish your task. The final image quality might be improved by slightly eroding the map to tighten it to the contours of the bird. You then also use the mask to calculate your color histogram by only taking foreground pixels into account. Edit: Look here:

Eroded mask

Final image

edited Sep 06 '18 at 16:58

answered Sep 04 '18 at 13:55

SilverMonkey

1,003
7
16

I am reading OpenCV and starting from step-1 , can you help me with some sample code to start with . – Sumeet Kumar Yadav Sep 04 '18 at 15:50
I personally do not use OpenCV, so i could only gather the respective methods for each step. – SilverMonkey Sep 06 '18 at 16:54
1

Pretty good answer to an under-specified question I think. Bearing in mind that identifying and segmenting a bird in a photo is very, very difficult, so this method pretty much takes the histogram of whatever object(s) are in focus, bird or not. If that's not good enough for the OP, they need to rephrase the question. – Chungzuwalla Sep 07 '18 at 00:50
This explanation is nice, I am reading opencv and trying now . If some one knows opencv can post with code – Sumeet Kumar Yadav Sep 07 '18 at 04:18

ElConrado · Answer 2 · 2018-09-07T16:36:07.353

According to this article https://www.pyimagesearch.com/2016/04/11/finding-extreme-points-in-contours-with-opencv/ and this question CV - Extract differences between two images

I wrote some python code as below. As my predecessor said it is also far from perfect. The main disadvantages of this code are constants value to set manually: minThres (50), maxThres(100), dilate iteration count and erode iteration count.

import cv2
import numpy as np

windowName = "Edges"
pictureRaw = cv2.imread("bird.jpg")

## set to gray
pictureGray = cv2.cvtColor(pictureRaw,  cv2.COLOR_BGR2GRAY)

## blur
pictureGaussian = cv2.GaussianBlur(pictureGray, (7,7), 0)

## canny edge detector - you must specify threshold values
pictureCanny = cv2.Canny(pictureGaussian, 50, 100)

## perform a series of erosions + dilations to remove any small regions of noise
pictureDilate = cv2.dilate(pictureCanny, None, iterations=20)
pictureErode = cv2.erode(pictureDilate, None, iterations=5)

## find the nozero regions in the erode
imask2 = pictureErode>0

## create a Mat like pictureRaw
canvas = np.full_like(pictureRaw, np.array([255,0,0]), dtype=np.uint8)

## set mask 
canvas[imask2] = pictureRaw[imask2]
cv2.imwrite("result.png", canvas)

Feature extraction and take color histogram

2 Answers2