Anntotate episodic events of videos
Project description
Behavior Annotator
Table of contents
Installation
- Install Git
- In git bash, run
git clone https://github.com/hsingchien/Bannotator.git
- In anaconda prompt, go to annotator directory, run
conda env create -f environment.yml
- Then activate the environment
conda activate bannotator
- Under this environment, go to the bannotator directory, run
python setup.py develop
- To open the app, under this environment, run
annotate-behavior
Usage
GUI
The GUI layout has 5 major areas.
- Video display area, where you can add multiple videos.
- Stream area shows the enlarged portion of the color coded streams.
- Stream overviews provide the bird's-eye view of the full annotation and the current progress.
- Behavior table dock where ID, name, keystroke and color are shown in the Behaviors tab; number of epochs and percentage of the time are shown in the Stats tab.
- Epoch tables where epochs are shown in the All Epochs tab, epochs of selected behaviors are shown in the Behavior Epochs tab.
Video
-
Annotator supports major video formats and the Norpix Streampix seq files. Only uncompressed seq (RAW) or compressed seq in JPEG format are currently supported.
-
Annotator is capable of displaying multiple videos simultaneously, therefore is great for experiments with multiple video streams. If videos have different number of frames, the 1st added video is treated as the main video and the other videos are stretched to the length of the video 1.
-
Annotator provides several different video layouts. Side by side, Stacked or Grid (this option becomes available when number of videos reaches 4).
-
Video can be played at various speed.
Annotation
The annotator is compatible with the format of annotation txt file of Piotr's MATLAB toolbox. If you are a user of the behavior annotator of this toolbox, you can view/edit your existing annotation files as well as keep using your configuration files to create new annotations. The output annotation txt files are also back compatible with Piotr's annotator.
If you do not have configuration files to start with, you can create an new annotation from scratch using New annotaion
dialog in the Annotation
menu. To enable this option you need to add a video first so that the annotator knows the length of the annotation.
First, set the number of streams. In the text editor, input behavior - keystroke pair in each line, parsed by '-' or space. The dialog actively checks the input to ensure the behavior-keystroke pairs are unique. Once created, all the streams will be initialized with the first behavior in your list. Usually the first behavior label is reserved for the blank label of name "other" or "blank", which will be assigned grey color.
Once the new streams are created, go through the video and label the events. Press keystroke to label the current frame as well as the rest of the Epoch.
Epoch
is the fundamental unit of the annotation. An Annotation
contains Streams
which contains a series of Epoch
. Each Stream
also contains several Behavior
objects each one of which stores name, color, keystroke and ID of a user defined behavior. Behavior
objects also collects all the Epochs
of this kind in its stream.
KeyPress Functions
Space : Play/Pause video
↑ / ↓ : Increase/Decrease playing speed
← / → : Previous/Next frame
1 - 0 : Change the current stream to stream 1 - 10
` : Rotate current stream through all streams
-/+ : Move to previous/next epoch of current stream
CTRL-/CTRL+ : Move to previous/next epoch of selected behavior
Contributing
Your contributions are always welcome! Contribution guideline will be available soon.
Acknowledgement
This project is inspired by Piotr's MATLAB toolbox.
Seq file reading is inspired by PIMS.
Random color generator distinctipy
The project is built with PySide6.
License
The MIT License (MIT) 2017 - Xingjian Zhang.
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
File details
Details for the file bannotator-1.0.0.tar.gz
.
File metadata
- Download URL: bannotator-1.0.0.tar.gz
- Upload date:
- Size: 1.1 MB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/4.0.2 CPython/3.9.16
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | 62e5d5c570ed5486a316dbb14f403fe75ddca8c388c99bf89974e587ff0a8f29 |
|
MD5 | dac7110ee13c1b6585201704484324c3 |
|
BLAKE2b-256 | 9ce2f2c37d60d63045ad4c12865fab6c771002cf85ce7f2a1b7b68c88bd5d189 |
File details
Details for the file bannotator-1.0.0-py3-none-any.whl
.
File metadata
- Download URL: bannotator-1.0.0-py3-none-any.whl
- Upload date:
- Size: 1.2 MB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/4.0.2 CPython/3.9.16
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | e8693fa47ff3f2481e34892aabcce27753af55036b850d36af8e5e687c16a420 |
|
MD5 | 5ed66e9ecb24daed25a3e9634bc29a20 |
|
BLAKE2b-256 | b5677696f1a19d73c0fe220e32303f0f2d93f17a6e2608ee0f9d05aeba257481 |