etikέta* is a cutting-edge web application designed to simplify and enhance the process of creating, training, and deploying AI models for object detection and segmentation. With a strong emphasis on georeferenced imagery, etikέta is perfect for projects that require spatial context, from environmental science and urban planning to defense and security. Developed by Scidrones, a leader in coastal mapping, remote sensing, and geoinformatics, etikέta leverages the latest in AI and geospatial technologies — including GeoAI techniques — to provide a user-friendly platform for data labeling, AI model training preparation, and inference evaluation. etikέta supports a wide range of applications, contributing to sustainable development and security initiatives.
*(eh-tee-KEH-ta) — from the Greek ετικέτα, meaning “label” or “tag”
Traditional methods of data labeling and AI training can be time-consuming and lack the spatial context needed for many applications. etikέta addresses these challenges by:
- Providing a collaborative platform for team-based project management
- Offering advanced labeling tools, including support for georeferenced images and automated assistance with tools like Segment Anything
- Enabling seamless export of labeled datasets in YOLO format for immediate use in AI training
- Facilitating model refinement through an intuitive inference mode
With etikέta, users can accelerate their AI development cycles while ensuring the highest accuracy and spatial relevance.
- Project Collaboration: Easily create projects, add team members, and define custom classes for your AI models.
- Advanced Labeling Tools: Upload your images, annotate object masks with precision, and utilize state-of-the-art tools like Segment Anything to accelerate your labeling workflow. For georeferenced images, your labels inherit spatial coordinates, enabling powerful geospatial analysis.
- Seamless Data Export: Export your labeled datasets in YOLO format, with optional tiling, ready for immediate use in training your AI models. Download georeferenced objects where applicable.
- Intelligent Inference: In inference mode, review AI-generated detections, mark images for retraining, and refine your models. Easily integrate your results with other systems for a complete data pipeline.
etikέta operates in two main modes: Labeling Mode and Inference Mode.
- Labeling Mode: Upload photographs or select areas on οn a world map using WMS services to generate new images for labeling. Annotate object masks for specific classes. If the images are georeferenced, the masks inherit spatial coordinates. Use tools like Segment Anything to assist in labeling, speeding up the process. Export the labeled dataset in YOLO format for AI training. In case the photographs are georeferenced the objects are exported in a geospatial format as well.
- Inference Mode: Connect your database to etikέta (dataset, classes, objects). Review and edit AI detections, mark images for retraining, and send refined results to other pipelines.
This straightforward workflow allows users to manage time-consuming parts of the AI training process within a single platform, from data preparation to model refinement.
Scidrones is a pioneering spin-off from the University of the Aegean, officially established in 2022. Headquartered on Lesvos Island, Greece, the company was founded in 2021 by a team of experts specializing in remote sensing, geoinformatics, artificial intelligence (AI), and Unmanned Aerial System (UAS) technologies.
Built on over a decade of cutting-edge research, Scidrones transforms complex geospatial data into actionable insights. Our advanced analytical tools and state-of-the-art sensor technology enable precise mapping, monitoring, and quantification of emerging phenomena across diverse industries.
At Scidrones, we bridge the gap between research and real-world applications, empowering organizations with data-driven decision-making solutions that enhance efficiency, accuracy, and innovation.
Tel: +30 694 429 8291
Email: info@scidrones.com
Address: Evagelistrias 30, Akleidiou, 81100 Mytilene, Greece