Deep Learning Image Labeling Tool Summary

For supervised learning algorithms, the data determines the upper limit of the task, and the algorithm just keeps approaching the upper limit. The furthest distance in the world is that we use the same model, but we have different tasks. But data labeling is a time-consuming effort, and here are a few image labeling tools: LabelMe

LabelMe data set for image segmentation tasks:

It comes from the following project: Https://github.com/wkentaro/labelme
This software realizes the most basic partition data labeling work, after save will keep some information of object into a JSON file, as follows:
Https://github.com/wkentaro/labelme/blob/master/static/apc2016_obj3.json
The software also provides the ability to convert the JSON file to Labelimage:
labelimg

LabelMe data set for image inspection tasks:

It comes from the following project: HTTPS://GITHUB.COM/TZUTALIN/LABELIMG
The label storage function and the "Next image", "Prev image" design is more convenient to use.
The last XML file format saved by the software is the same as the Imagenet dataset. Yolo_mark

Yolo_mark data set for image inspection tasks:

It comes from the following items: Https://github.com/AlexeyAB/Yolo_mark
It is a Yolo2 team open-source image labeling tool that allows others to use Yolo2 to train their task models. Both Linux and win can be run, relying on the OpenCV library. vatic

Vatic data set for image inspection tasks:

It comes from the following project: http://carlvondrick.com/vatic/
In particular, it can do video annotations, such as a 25fps video, only need to manually mark the position of the object around 100 frames, and finally in the entire video can have a better effect. This relies on a software-integrated OPENCV tracking algorithm. Sloth

Sloth data set for image inspection tasks:

It comes from the following items:
Https://github.com/cvhciKIT/sloth
https://cvhci.anthropomatik.kit.edu/~baeuml/projects/a-universal-labeling-tool-for-computer-vision-sloth/
When labeling a label, the software can store the label and render a list of bbox in the highlighted picture. annotorious

Annotorious data set for image inspection tasks:

It comes from the following items:
Http://annotorious.github.io/index.html
Code written in a fairly canonical, providing the corresponding API interface, easy to directly modify and invoke. Rectlabel

Rectlabel data set for image inspection tasks:

It comes from the following items:
https://rectlabel.com/
This is a software for Mac OS X and can be downloaded directly from the Apple App Store. Vott

Vott data set for image inspection tasks:

It comes from the following items:
https://github.com/Microsoft/VoTT/
Microsoft's Open source tools, which can annotate video or annotate images, and support the integration of existing models, are powerful. iat–image Annotation Tool

IAT for the creation of data sets for image segmentation tasks:

It comes from the following items:
http://www.ivl.disco.unimib.it/activities/imgann/
What is more distinctive is that it supports some basic shape selection, such as the object to be segmented is a circle, then the division can be directly select the circle, instead of using the polygon point. Images_annotation_programme

Images_annotation_programme data set for image inspection tasks:

It comes from the following items:
Https://github.com/frederictost/images_annotation_programme

Web version of OH

In addition, there are a lot of similar tools, compared with the above tools, and there is no feature, we only give links, not detailed description: imagenet-utils

Https://github.com/tzutalin/ImageNet_Utils Labeld

Https://github.com/sweppner/labeld VIA

http://www.robots.ox.ac.uk/~vgg/software/via/ ALT

https://alpslabel.wordpress.com/2017/01/26/alt/ Fastannotationtool

Https://github.com/christopher5106/FastAnnotationTool Lera

Https://lear.inrialpes.fr/people/klaeser/software_image_annotation