I've loaded an image using CV2.imread, and saved it using cv2.imwrite() and scipy.misc.imsave(). In both these cases output image size is bumped. Why is this?
Both input and output images are of file type .jpg
img = cv2.imread(img_src)
scipy.misc.imsave(img, "scipy_original.jpg")
cv2.imwrite("cv2_original.jpg", img)
Input file size is 309kb
Output file size in cv2 is 690kb
output file size in scipy is 399kb
this is the image if you want reference: https://i.imgur.com/0J8ClQn.jpg
OpenCV has different compression levels for jpg, from 0 to 100. The default is 95.
This question discusses it, and this site has examples.
Saving with parameters like this:
cv2.imwrite("cv2_original.jpg", img, [int(cv2.IMWRITE_JPEG_QUALITY), jpg_quality]) where jpg_quality < 95 will reduce the file size.
Related
I am using GIMP 2.10.24. I have some image and I need to change Print Size Width to 21mm and Height to 30mm.
I can do that with Set Image Print Resolution Dialog (Menu->Image->Print Size):
screenshot
But there is my question: how could I do that using script-fu or python-fu?
Print size, size in pixels, and print definition are completely related:
print size = size in pixels ÷ print definition
So to change the image print definition you use
In Python:
pdb.gimp_image_set_resolution(image, xresolution, yresolution)
In Script-fu:
(gimp-image-set-resolution image xresolution yresolution)
In both case the X/Y resolutions are in dots per inch.
However if you are using Gimp just for this creating a Gimp script is overkill (the learning curve is quite steep). If the image is in a common format (JPEG, PNG, TIFF) the print definition is part of the image metadata (JPEG header, or EXIF data) and can be changed directly without decoding/reencoding the image using CLI utilities. For instance with ExifTool:
exiftool ${your_image} -xResolution=321 -yResolution=321
I am using opencv in python to rotate an image and the original and the resulted images are differrent is somethings, I am doing my transformation through this part of code:
img = cv2.imread("image.tif")
new_image = cv2.getRotationMatrix2D((cols / 2, rows / 2), correction_angle, 1)
dst = cv2.warpAffine(img, new_image , (cols, rows))
cv2.imwrite("Rotated_image.tif", dst)
The original image's size is 1.7 Mb, The image's resolution is 300
dpi, and the color space is YCbCr.
The issue is that the resulting image with 12.5 Mb size, 96 dpi, the color space is RGB, and with compression "LZW"!
My question is that: Can I keep the main properties of the original image? and why rotating an image changes the size this way?
Note: The bit depth is 24 in both images.
Calling cv2.imread with only the name of the file uses the default value cv.IMREAD_COLOR for the second parameter, about which the documentation says:
If set, always convert image to the 3 channel BGR color image.
So, your image is always converted to RGB. You can try using cv.IMREAD_ANYCOLOR for the second parameter of imread, but I don't think you can use cv2.warpAffine on it trivially then.
The difference in the stored DPI information stems from the fact that you write the new image without any meta data. imwrite allows you to specify parameters (see here), but, unfortunately, they are not very well documented. I am not sure if this kind of data can be written out of the box with OpenCV.
I am trying to read a medical imaging data which is in .tif format using Octave . It so happens that only software from https://fiji.sc/ has the ability to read the medical images I have. A sample of the image that I am referring is this one. This image is readable in.tif format only in fiji. The image can also be converted to jpg image. Only after data points is visible to other image editing software. Now the issue is , I want to load it into octave and read it - at the moment it is not possible. The imshow() function in octave is not allowing me to visualize the given .tif file.
So, if anybody has experience in reading such file formats please let me know. Thanks.
Your linked image is a 16bit grayscale image, see $ gm identify -verbose hello.tif. So lets load it into GNU Octave:
img = imread ("hello.tif");
hist (img(:), 200); # show histogram
Ah!, the main information of your image is in the range 0-600 (probably the image aquisition system used had 11 or 12bit resolution), so lets scale and print this part as color encoded (viridis) image:
imagesc (img, [0 600])
colorbar
Is this what you want? Of course we can also use a gray colormap, try: colormap gray
If this still doesn't fit your need you should really explain what you expect to see....
imread() relies on image magic to do the conversion, some versions of which can only read TIFF in 8-bit mode (giving a warning message), but when and how this happens I do not know.
It works for me using a .raw ṕicture converted to .tif in Fiji. I am specifically interested in several spectral analysis from my pictures which are taken illuminating with 660 and 850nm.
I am automating conversion of source PNG images to JPEGs of a predefined dimension. For most the images, I don't need to provide the sampling-factor and happy with the output quality and file size. However, a few of the files get heavily distorted with artifacts. For such files, I currently manually provide the option '-sampling-factor 1x1' to get the desired output jpeg, though bigger file size.
Is there an way to identify before hand which PNG src file needs the usage of sampling-factor for conversion? That will help to pull it in the script.
I need to edit a png image,by giving it border and drop shadow effect. But the final size of the edited image is too high to use for a mobile app .I know that size of jpeg is less compared to that of png.So i convert that image to jpeg and tried to give drop shadow and border effect.But that image is not having transparent background..Is their any other methods to accomplish this using jpeg?
Another option is to try either ImageOptim for losseless compression, or its lossy cousin, ImageAlpha.
ImageOptim tries a series of lossless algorithms to shrink a PNG and selects the smallest result of the bunch. It has taken 25% to 50%+ of quite a few of our images.
ImageAlpha, on the other hand, is lossy and can further crunch the image, with results more like JPEG but without losing Alpha.
You would also do well to disable PNG compression in Xcode as shown here, with additional details here.
What #minitech wanted to say is not about scaling, it's about file compression. jpg and png files usually have some data that could be removed from the file. There are some compression methods to reduce file size (note that is size in kbs, not in scale measurement). Jpg images can reduce file size by reducing image quality, too.
If you want another file type that accepts transparency, there are the gif format, which gives you a smaller file, but have some drawbacks, like a lack of alpha channel (variable transparency). Check this link for more details: http://www.w3.org/QA/Tips/png-gif
There are a couple of online file compressors. If you want to compress png files, you could try using http://tinypng.org/
No, jpeg image wont support transparency.But you can change the white background coming along with jpeg image