如何仅使用标准库以Python读取任意图像文件格式（PNG，JPEG，TIFF，BMP）？

Question

I'm an experienced Python programmer with plenty of image manipulation and computer vision experience. I'm very familiar with all of the standard tools like PIL , Pillow , opencv , numpy , and scikit-image .

How would I go about reading an image into a Python data format like a nested list, bytearray, or similar, if I only had the standard library to work with?

I realize that different image formats have different specifications. My question is how I would even begin to build a function that reads any given format.

NOTE Python 2.6 had a jpeg module in the standard library that has since been deprecated. Let's not discuss that since it is unsupported.

Answer 1

If you're asking how to implement these formats "from scratch" (since the standard libraries don't do this), then a good starting point would be the format specification.

For PNG, this is https://www.w3.org/TR/2003/REC-PNG-20031110/ . It defines the makeup of a PNG stream, consisting of the signature (eight bytes, 8950 4e47 0d0a 1a0a , which identifies the file as a PNG image) and a number of data chunks that contain meta data, palette information and the image itself. (It's certainly a substantial project to take on, if you really don't want to use the existing libraries, but not overly so.)

For BMP, it's a bit easier since the file already contains the uncompressed pixel data and you only need to know how to find the size and offset; some of the format definition is on Wikipedia ( https://en.wikipedia.org/wiki/BMP_file_format ) and here: http://www.digicamsoft.com/bmp/bmp.html

JPG is much trickier. The file doesn't store pixels, but rather "wavelets" which are transformed into the pixel map you see on the screen. To read this format, you'll need to implement this transformation function.