Python implementation of data using simple annotation tools
In order to increase efficiency for the purpose, we have recently been working to build an AI ToolBox
These two days, I've added data annotation module in which pretreatment tool catalog, this article describes the contents of the code, see here
The module contains the following data labeling demo
gui_tkinter_exercise.py
data_annotation_tool.py
image_select_tool.py
A simple GUI exercise
gui_tkinter_exercise.py
Is a simple GUI exercise demo, which shows the basic usage of the python Tkinter GUI library
If you are not familiar with Tkinter, reading gui_tkinter_exercise.py
code is a very good choice, you will learn how to create a GUI form, how to create a Button, how to correlate events, and create a knowledge annotation tools necessary data.
The GUI demo is shown as follows:
Data Highlighter
data_annotation_tool.py
It is a data tagging demo, shows how to implement a data annotation tool
how to use?
First, the text box shows the path to save data marked results (currently for convenience is hard-coded in the code)
Then, click on the 'start' button to select the folder where the data to be annotated
After that, the specified folder of all pictures will be loaded one by one in sequence
Click the Categories button (eg 'Dog' and 'Cat') to annotate a picture
Click on 'NumericUpDown' button to jump to a specific picture marked the beginning (before work can be used to continue)
Click on 'QUIT' to quit
Data Highlighter run shot as follows:
Data filtering tools
image_select_tool.py
Data filtering is a demo, shows how to implement a data screening tool. It is a tool by the actual needs of the birth of my work
Potential scenarios:
Suppose we have a lot of sets of images that we want to filter out one or more pictures from each group. When you have such similar needs, this tool suited to help you improve efficiency
Run the program logic is:
First, the text box shows the path to save data marked results (currently for convenience is hard-coded in the code)
Then, click on the 'start' button to be screened to select the root folder where data
Subsequently, all subfolders under the root folder of the folder will be loaded one by one
Each time, a single sub-folder of all images will be presented in the canvas Canvas
Use the mouse to click on the image you want to filter out, it will be automatically saved to the specified path
Use the mouse wheel to quickly browse images
To press the space switch to the next group of pictures
Use screenshot of the tool is as follows: