Click the above“Beginner Learning Vision” to select “Star” or “Top”

Important content delivered promptly

For academic sharing only, does not represent the stance of this public account. Contact for deletion if infringement occurs.

Reprinted from: Author丨Dmitrii

LabelMe

Address: https://github.com/wkentaro/labelme

What can you do with it?

LabelMe is an open-source image polygon annotation tool based on Python, which can be used for manual image annotation for object detection, segmentation, and classification. It is the offline branch of the online LabelMe, which has recently closed new user registration options. Therefore, in this article, we only consider LabelMe (in lowercase).

The tool is a lightweight graphical application with an intuitive user interface. With LabelMe, you can create: polygons, rectangles, circles, lines, points, or polylines. Generally, it is convenient to export annotations in well-known formats (such as COCO, YOLO, or PASCAL VOC) for later use. However, in LabelMe, labels can only be saved directly from the application as JSON files. If you want to use other formats, you can use the Python scripts in the LabelMe repository to convert the annotations to PASCAL VOC. Nonetheless, it is still a fairly reliable application with simple features for manual image labeling and a wide range of computer vision tasks.

Installation and Configuration

LabelMe is a cross-platform application that works on multiple systems such as Windows, Ubuntu, or macOS. The installation itself is very simple, and there is a good description here. For example, on macOS, you need to run the following commands in the terminal:

1. Install dependencies: brew install pyqt
2. Install LabelMe: pip install labelme
3. Run LabelMe: labelme

LabelImg

Address: https://github.com/tzutalin/labelImg

What can you do with it?

LabelImg is a widely used open-source graphical annotation tool. It is only suitable for object localization or detection tasks and can only create rectangular boxes around the objects of interest. Despite this limitation, we still recommend using this tool because the application focuses solely on creating boundary boxes as simply as possible. For this task, LabelImg has all the necessary features and convenient keyboard shortcuts. Another advantage is that you can save/load annotations in three popular formats: PASCAL VOC, YOLO, and CreateML.

Installation and Configuration

Here is a good description of the installation. Also, note that LabelImg is a cross-platform application. For macOS, you need to execute the following on the command line:

1. Install dependencies: first brew install qt, then brew install libxml2
2. Select the folder location for installation.
3. When you are in the folder, run the following commands: git clone https://github.com/tzutalin/labelImg.git, cd labelImg then make qt5py3
4. Run LabelImg: python3 labelImg.py
5. Developers strongly recommend using Python 3 or higher and PyQt5.

CVAT

Address: https://github.com/openvinotoolkit/cvat

What can you do with it?

CVAT is an open-source annotation tool for images and videos, used for tasks such as object detection, segmentation, and classification. To use this tool, you do not need to install the application on your computer. You can use the web version of this tool online. You can collaborate as a team to process annotated images and assign work among users. There is also a great option that allows you to use pre-trained models to automatically annotate your data, which can simplify the process for the most common classes (such as those included in COCO) if you use the existing models available in the CVAT dashboard. Alternatively, you can also use your own pre-trained models. CVAT has the widest feature set among the tools we have considered. In particular, it allows you to save labels in about 15 different formats. A complete list of formats can be found here.

Hasty.ai

Address: https://hasty.ai/

What can you do with it?

Unlike all the tools mentioned above, Hasty.ai is not a free open-source service, but it is very convenient for labeling data due to its AI assistant for object detection and segmentation. The automatic support allows you to significantly speed up the annotation process, as the assisting model is being trained during the labeling. In other words, the more labeled images you have, the more accurate the assistant’s work becomes. We will look at an example below to illustrate how it works.

You can also try this service for free. The trial offers 3000 credits, enough to automatically generate approximately 3000 suggested labels for an object detection task.

Hasty.ai allows you to export data in COCO or Pascal VOC format. You can also work on a single project as a team and assign roles in the project settings.

Once the free credits are used up, Hasty.ai can still be used for free, but the labeling will be entirely manual. In this case, it is better to consider the free tools mentioned above.

Configuration

1. To use the tool, you need to register on hasty.ai.
2. Log into your account.
3. Click to create a new project.
4. Fill out the form with a name and description and navigate to the project settings, where you can define the classes to consider and add data to the project.
5. Additionally, you can add other users to collaborate on the project. Credits will be used from the accounts of users sharing the project.

Original article link: https://medium.com/dida-machine-learning/the-best-labeling-tools-for-computer-vision-bf4a9642f796

——————-

END

——————–

Download 1: OpenCV-Contrib Extension Module Chinese Tutorial

Reply "Extension Module Chinese Tutorial" in the backend of "Beginner Learning Vision" public account to download the first Chinese version of the OpenCV extension module tutorial available online, covering installation of extension modules, SFM algorithms, stereo vision, object tracking, biological vision, super-resolution processing, and more than twenty chapters of content.

Download 2: Python Vision Practical Project 52 Lectures

Reply "Python Vision Practical Project" in the backend of "Beginner Learning Vision" public account to download 31 visual practical projects including image segmentation, mask detection, lane line detection, vehicle counting, eyeliner addition, license plate recognition, character recognition, emotion detection, text content extraction, facial recognition, etc., to help quickly learn computer vision.

Download 3: OpenCV Practical Project 20 Lectures

Reply "OpenCV Practical Project 20 Lectures" in the backend of "Beginner Learning Vision" public account to download 20 practical projects based on OpenCV, achieving advanced learning of OpenCV.

Group Chat

Welcome to join the public account reader group to communicate with peers. There are currently WeChat groups for SLAM, 3D vision, sensors, autonomous driving, computational photography, detection, segmentation, recognition, medical imaging, GAN, algorithm competitions, etc. (will gradually be subdivided in the future). Please scan the WeChat number below to join the group, and note: "Nickname + School/Company + Research Direction", for example: "Zhang San + Shanghai Jiao Tong University + Vision SLAM". Please follow the format for remarks, otherwise, you will not be approved. After successful addition, you will be invited to join the relevant WeChat group based on your research direction. Please do not send advertisements in the group, otherwise you will be removed from the group, thank you for your understanding~