Labelo provides robust import and export functionalities that streamline the process of managing datasets for annotation projects. These features support a variety of data types, including images, text, videos, and audio, ensuring versatility in handling different annotation tasks. Because of this, labelo can be used as,

1. Image labeling tool
2. Video labeling tool
3. Text labeling tool
4. Audio tagging software

Importing Data

Types of Data Supported:

• Images: Upload image files such as JPEG, PNG, or TIFF for annotation.
• Text: Import text documents for tasks like text classification or named entity recognition.
• Videos: Upload video files to annotate frames or segments.
• Audio: Import audio files for tasks like speech recognition or sound classification.

Import Methods:

• Direct Upload: Use the web interface to upload files directly from your computer. You can drag and drop files or browse to select them.
• Cloud Storage Integration: Connect to cloud storage services like AWS S3, Google Cloud Storage, or Azure Blob Storage. This allows you to import large datasets stored in the cloud without downloading them locally.
• URLs: Import data from publicly accessible URLs. This method is useful for web scraping or using online datasets.

Import Workflow

We can import datasets in the open-source labeling tool in two ways, either from the project page itself while creating a project or from the task page of each project by clicking the import button.

Labelo’s data import modal streamlines the process of setting up new annotation projects by integrating multiple data import options directly within the project creation workflow. Importing data is made convenient through several methods available right within the modal. 

It includes a drag-and-drop feature that significantly simplifies the process of importing data from your computer. This functionality allows you to quickly add various types of files by simply dragging them from your local file system and dropping them directly into the designated area within the modal. This method is particularly useful for quickly adding a diverse range of data types, including images (such as JPEG, PNG, or TIFF files), text documents (like TXT or CSV files), and other formats used in annotation projects.

In addition to the drag-and-drop feature, Labelo’s data import modal supports traditional file uploads, providing another convenient method for importing data into your project. This functionality allows you to manually browse and select files from your local storage, which can be especially useful if you prefer a more controlled and precise way to choose specific files for your annotation project. To use this feature, you simply click on the “Upload” button within the modal, which opens a file dialog window. From here, you can navigate through your computer’s file system to locate and select the files you wish to import. This method supports the selection of multiple files at once, enabling you to add several files in a single action. Additionally, it accommodates various data types, including images, text documents, videos, and audio files, ensuring compatibility with diverse project requirements.

For users who have datasets hosted online, Labelo’s data import modal offers a convenient option to add URLs, allowing you to import data directly from web sources without the need to download it first. This feature is particularly beneficial for those who work with large datasets or datasets that are frequently updated and stored on external servers. For users who have datasets hosted online, Labelo’s create project modal offers a convenient option to add URLs, allowing you to import data directly from web sources without the need to download it first. This feature is particularly beneficial for those who work with large datasets or datasets that are frequently updated and stored on external servers.

Exporting Datasets

Once annotations are completed, Labelo offers multiple export options tailored to various needs and applications. These options ensure that your annotated data can be seamlessly integrated into different machine-learning frameworks and workflows. Here’s an elaboration on each export format:

COCO (Common Objects in Context)

COCO is a widely adopted format in the computer vision community, ideal for tasks involving object detection, segmentation, and keypoint detection. This format is known for its rich metadata, which includes detailed information about each image, such as object categories, bounding box coordinates, segmentation masks, and key points. The comprehensive nature of COCO annotations makes it suitable for training complex models that require detailed contextual information. Additionally, COCO is compatible with popular machine learning frameworks like TensorFlow, PyTorch, and Detectron2, facilitating seamless integration into various workflows and enhancing the dataset’s usability for advanced computer vision applications.

Pascal VOC (Visual Object Classes)

Pascal VOC is another prominent format used primarily for object detection tasks. The format organizes annotations in XML files that describe the objects within each image, including their classes and bounding box coordinates. The XML-based structure is straightforward and easy to parse, making it a popular choice for many computer vision projects. Pascal VOC is supported by numerous machine learning libraries and frameworks, which enhances its applicability and ease of use. This format is particularly favored for its simplicity and effectiveness in handling standard object detection tasks, enabling efficient model training and evaluation.

CSV (Comma-Separated Values)

CSV is a versatile and widely used format suitable for a variety of data types, including text classification, tabular data annotation, and other non-image tasks. Its simple, human-readable structure makes it easy to work with and integrate into different applications. CSV files are often used for basic data analysis and are compatible with many software tools, including spreadsheet programs like Excel and programming languages like Python. The format’s flexibility allows it to be used in diverse scenarios, from data annotation to model training, making it an essential tool for handling structured data in a straightforward manner. Discover more about Labelo’s team and operations by exploring our detailed insights into the role of the Data Manager and its significance within the organization. Learn more about the critical responsibilities and contributions by reading our article, What is the Role of Data Manager in Labelo.