pixels2svg 0.1.0

pixels2svg : Convert pixels to SVG square-based shapes.

pixels2svg _^v0.1.0

Convert pixels to SVG square-based shapes.

There are a few open source tools around that do bitmap -> vector tracing, for instance potrace which is the engine used by Inkscape and also has available python bindings (pypotrace).

But sometimes, what you want is the exact bitmap data, but as a set of SVG shapes, that you could for instance:

• enlarge for big prints
• use as input paths for a wood laser cutter
• customize in a vector graphics software (edit opacity, contours, overlay on some background image etc.)

Medical image segmentation is another typical use case where one has pretty low-res images and sometimes needs to display them best for data visualization/presentation purposes. In this case, an SVG overlay with editable contours and opacity can be convenient.

This package enables you to easily convert your bitmap image to its equivalent SVG representation, each path of the SVG being a polygon made of adjacent pixels of same color merged together.

It is inspired of Florian Berger's pixel2svg package (GitHub fork available here), which is pretty old (python2) and doesn't merge adjacent pixels of same color, making the SVG pretty tedious to edit in vector graphics softwares.

Example 1 - pixel art

File	Preview
Original PNG image (32×32)
Original PNG image (browser enlarged)
Output SVG image
Output SVG image with customized contour style (see examples/sword_outline.py)

Example 2 - brain scan + segmentation overlay

File	Preview
Scan image (PNG)
Segmentation overlay (PNG)
Converted SVG segmentation overlay
Converted SVG overlayed on PNG scan with customized opacity & contour style (see examples/sword_outline.py)

Requirements

• Python >= 3.7

---

Dependencies

• svgwrite
• PIL
• cc3d

---

Install

pip install pixels2svg

or directly from Github:

pip install git+git://github.com/ValentinFrancois/pixels2svg#egg=pixels2svg

---

Usage examples

CLI

usage: python3 -m pixels2svg [-h] [--output path] [--no_group_by_color] [--no_pretty] path

pixels2svg CLI

positional arguments:
  path                  input path of the bitmap image. If not passed, will print the output in the terminal.

optional arguments:
  -h, --help            show this help message and exit
  --output path, -o path
                        output path of the bitmap image
  --no_group_by_color   do not group shapes of same color together inside <g> tags
  --no_pretty           do not pretty-write the SVG code

In Python

Simple usage

from pixels2svg import pixels2svg

pixels2svg('input.png', 'output.svg')

pixels2svg() accepts a few optional arguments that should help you integrate it best within your codebase:

Parameters

• input_path : str
  Path of the input bitmap image

• output_path : Optional[str]
  Path of the output SVG image (optional). If passed, the function will return None. If not passed, the function will return the SVG data as a str or a Drawing depending on the as_string parameter.

• group_by_color : bool
  If True (default), group same-color shapes under SVG elements.

• as_string : bool
  If True and no output_path is passed, return a str representing the SVG data.

• pretty : bool
  If True (default), output SVG code is pretty-printed.

Returns

• Optional[Union[svg.Drawing, str]]
  Depends on the output_path and as_string parameters

Advanced usage

By setting output_path=None and as_string=False, you can get an object inheriting svgwrite.Drawing, with the additional useful methods:

• def save_to_path(self, path: str, pretty: bool = False)
• def to_string(self, pretty: bool = False) ‑> str

This gives you access to the svgwrite API to modify the output programmatically.

See advanced examples in examples.

---