https://github.com/jakubvalenta/image-marker
Mark rectangular areas in a batch of images and save the output as a CSV
https://github.com/jakubvalenta/image-marker
batch collection graphics image mark rectangle
Last synced: 10 months ago
JSON representation
Mark rectangular areas in a batch of images and save the output as a CSV
- Host: GitHub
- URL: https://github.com/jakubvalenta/image-marker
- Owner: jakubvalenta
- License: apache-2.0
- Created: 2017-09-10T17:22:54.000Z (over 8 years ago)
- Default Branch: master
- Last Pushed: 2020-07-11T15:08:30.000Z (over 5 years ago)
- Last Synced: 2025-01-30T11:11:40.285Z (12 months ago)
- Topics: batch, collection, graphics, image, mark, rectangle
- Language: Python
- Homepage:
- Size: 4.25 MB
- Stars: 0
- Watchers: 3
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Image Marker
A graphical program to mark rectangular areas in a batch of images and save
their coordinates as a CSV. Optionally, add a text note to each of the
images. Optimized for a fast workflow.
Output CSV format:
``` csv
```
Example:
``` csv
./test_images/building.jpg 350 237 909 547
./test_images/canal.jpg 530 533 1075 252
```
This is all Image Marker does. You can do whatever you want with the output
CSV. One obvious option is to crop the images according to the marks. You can
use [image-crop](https://github.com/jakubvalenta/image-crop) for that.
## Installation
``` shell
pip install --user --upgrade .
```
## Usage
1. Put all the images you want to mark in a single directory.
2. Run Image Marker:
``` shell
image-marker ./my_directory --marks my_marks.csv --output my_output.csv --box-ratio 1.33 --box-padding 0.1
```
The positional argument specifies the directory to load the images from.
The option `--marks` specifies the CSV file path, where the coordinates of
the marked rectangles will be saved.
The option `--output` specifies the CSV file path, where the post-processed
coordinates will be saved. More on that later.
3. A window will appear with the first image in the directory. Use the mouse to
draw a rectangular mark. The output CSV files are saved immediately.
If you're not happy with the mark position or size, draw it again. The
previously drawn rectangle will disappear.
4. Use the right arrow key to advance to the next image.
You can also go back using the left arrow key.
5. Type any characters on your keyboard to add a text note to an image. It will
appear in blue by the bottom edge of the window. You can delete what you've
typed using the backspace key.
6. Quit the program by pressing the esc key.
### Mark post-processing
By default the values stored in the `--marks` and `--output` CSV files are the
same.
However, it is possible to make some automatic adjustments to your marks. In
this case, the original marks will be stored in the `--marks` CSV file and the
post-processed marks in the `--output` CSV file.
There is currently one post-processing option:
#### Box
If you want to create a set of marks that all have the same aspect ratio, you
can use the option `--box-ratio` to automatically draw a box around each mark
with a specified aspect ratio. The box will be drawn symmetrically around the
marked area as a green rectangle. If the box doesn't fit in the image, it will
be moved in the image.
You can also add padding to the box using the option `--box-padding`. It is
defined as a fraction of the marked rectangle width. For instance `--box-padding
0.15` will result in a 30 px padding for a 200 px wide mark.
## Controls
- right or down or enter: show next image
- left orup: show previous image
- left mouse click and drag: draw a mark
- left mouse click (no drag): delete current mark
- esc: exit the program (all marks are automatically saved)
- letter and number keys: add text note to current image
- backspace: delete the last character of the text note
## Help
``` shell
image-marker --help
```
## Contributing
__Feel free to remix this piece of software.__ See [NOTICE](./NOTICE) and
[LICENSE](./LICENSE) for license information.