We are happy to announce the publication of embeddings derived from the MajorTOM dataset in collaboration with European Space Agency - ESA. The goal of this initiative is to make an open source dataset searchable with natural language - without the use of any metadata, labels or annotations. We aim to push the boundaries of ML technology and foster a spirit of collaboration and innovation within the open-source community. Our decision to open source the embeddings is driven by our desire to inspire community-driven exploration and development in large-scale vectorisation of visual datasets. We believe that by providing these resources freely, we can stimulate a wave of creativity and innovation in the field, leading to novel applications and advancements that we have yet to imagine. Furthermore, we are excited to see how the ML community leverages these embeddings for applications beyond our current scope. By fostering an environment of open collaboration and sharing, we aim to accelerate the advancement of earth observation technologies and contribute to the collective knowledge base of the computer vision and vector space community. If you have questions or feedback, please follow us on LinkedIn and Hugging Face and get in contact with us. https://lnkd.in/e5qhadFn #opensource #vectorembeddings #majortom #earthobsersation #vectors #vectorisation #dataintelligence #satelliteimagery #opensourcing #innovation #huggingface #esa #CLIP #siglip
Info
At Quasara, we believe that we need a strong infrastructure to build great enterprise applications on top of foundational models and GenAI. This is why we are on a mission to vectorise the world. Our first product is a vectorisation engine that vectorises visual data of enterprises efficiently, at scale and reliably. Our first use case is data exploration (EDA) for large video and image datasets used to develop applications for autonomous driving, infrastructure monitoring, surgeries and manufacturing. Our vectorisation engine can vectorise millions of images within 24 hours and make them searchable using natural language. At its core, this makes vast, unlabelled data accessible and gives it structure, semantic understanding and enables users to put it to use through dedicated data workflows. This enables companies to put vectors of their data at the center of their data and AI infrastructure.
- Website
-
https://www.quasara.io
Externer Link zu Quasara
- Branche
- Softwareentwicklung
- Größe
- 2–10 Beschäftigte
- Hauptsitz
- Munich, Bavaria
- Art
- Privatunternehmen
- Gegründet
- 2022
- Spezialgebiete
- Computer Vision, Data Curation, Data Quality, Unstructured Data, Data Quality Automation, Exploratory Data Analytics, Robot Vision, Machine Vision, Semantic Search Engine, Semantic Search, Vector Search, Vector Embeddings, Vectorisation und Vectorization
Produkte
Quasara Plato
Software für die Verarbeitung und Verteilung von Big Data
Our Data Intelligence Engine makes terabytes of raw images and videos searchable and usable. It provides an unparalleled semantic understanding – out-of-the-box – with zero pre-training or metadata needed. We use this semantic understanding to make data transparency, data management and data preparation of raw images and videos 10x easier than before. Get your new computer vision models from 0% to 80% within days. And save 70% of costs on data exploration and selection for labelling projects.
Orte
-
Primär
Prinzregentenstraße 54
Munich, Bavaria 80538, DE
-
Zurich, CH
-
Leipzig, DE
Beschäftigte von Quasara
Updates
-
Prompt of the week: “open pit mining” Let's see where in Germany we can find mining activities in the vast MajorTOM dataset. If you are interested in the embeddings we’ve created for the MajorTOM dataset, follow us and stay tuned. We are going to publish them shortly. A Big Thank You to the European Space Agency - ESA for sharing their satellite dataset publicly. Contact us if you want to vectorise your image or video data. Sources: https://lnkd.in/ezq-3-q4 https://lnkd.in/epqFQp2t #earthobservation #satelliteimages #vectorisationengine #dataintelligence #semanticsearch #vectorisation #dataexploration #computervision #datapreprocessing #dataexploration #openpitmining #mining #coalmine
-
Query of the week on Major-TOM: "airports" 🛫 🛬 . Let's search for some of Germany's airports in the vast MajorTOM dataset of European Space Agency - ESA. What we did: a) vectorising the entire Major-TOM image data for Germany; b) using our algorithms and models to make sure we also find the smallest parts of the satellite footage; c) no pre-training, no labelling happened in the preparation of this dataset. A Big Thank You to the European Space Agency - ESA for sharing their satellite dataset publicly, propelling our R&D. Contact us if you want to learn how #vectorisation and #dataexploration on unlabelled visual data helps you analyse, structure, and prepare satellite images and save time and money while doing it. Sources: https://lnkd.in/ezq-3-q4 https://lnkd.in/epqFQp2t #earthobservation #satelliteimages #vectorisationengine #dataintelligence #semanticsearch #vectorisation #dataexploration #computervision #datapreprocessing #dataexploration #airport Mikolaj Czerkawski
-
It's time to look at some satellite imagery of European Space Agency - ESA and their Major-TOM dataset. As the EM2024 is in full swing, let's search for some of the stadiums in Germany where our talents are discovered. What we did was a) to vectorise the entire Major-TOM image data for Germany; b) used our algorithms and models to make sure we also find the tiniest stadiums; and c) no pre-training, no labelling happened in the preparation of this dataset. A Big Thank You to European Space Agency - ESA for sharing their satellite dataset publicly, enabling our R&D on satellite imagery. Contact us and follow us if you want to learn how #vectorisation and #dataexploration on unlabelled satellite images helps you analyse, structure and prepare data and save money while doing it. Sources: https://lnkd.in/ezq-3-q4 https://lnkd.in/epqFQp2t #earthobservation #satelligeimages #vectorisationengine #dataintelligence #semanticsearch #vectorisation #dataexploration #computervision #datapreprocessing #dataexploration
-
Experience smart recording with Quasara Plato „Intelligence Anywhere“ solution in partnership with b-plus Group at our booth 6128 today. Let‘s have a chat! ADAS & Autonomous Vehicle Technology Expo #autonomousdriving #ADAS #smartrecording #AI #GenAI #efficiency
You might have spotted b-plus not only on stand 6126 at AVTE in Stuttgart but also with many partners. Adrian spotted our AVETO.dps at Quasara‘s stand just next to us and Christoph could tell him more about their application. Quasara's AI-powered smart recording triggering technology optimizes data collection together with the b-plus processing and recording solution AVETO.dps. Our innovative combination reduces unnecessary data volumes and saves storage space by recognizing and recording relevant events selectively and precisely. It also enables flexible object recognition that adapts flexibly and dynamically to your individual needs with voice command. Quasara's zero-shot AI enables immediate and precise detection of any amount of objects, situations or edge cases. Visit us and Quasara‘s stand to see the live demo and get your personal experience of the smart recording trigger. ADAS & Autonomous Vehicle Technology Expo #adas #autonomousdriving #realtimeanalytics #smartrecording #genai #ai #datarecording
-
Prompt of the Week: "Woman pushing a stroller." Experience the power of vectors of visual data when combined with natural language prompting. It enables #DataExploration on unlabelled image or video data. Special thanks to AUDI AG for open sourcing their groundbreaking dataset, a2d2, driving our research forward. Sources: a2d2.audi https://lnkd.in/eUXSn_tU https://lnkd.in/eAVExMDG #Imagevectorisation #Vectorisation #VectorSpace #AutonomousDriving #ADAS #DataIntelligence #SemanticSearch #UnderstandRawData #DataAcquisition #ComputerVision #ScenarioCatalogue #DataPreprocessing #DataExploration
-
Prompt of the Week: "Ice Cream" The warm weather calls for some delicious ice cream and we found the biggest ice cream on the road 🍦 Witness Quasara Plato's innovative approach to uncovering insights in unlabelled datasets through natural language processing. Discover the power of #DataExploration on unlabelled image or video data. Special thanks to AUDI AG for open sourcing their groundbreaking dataset, a2d2, driving our research forward. Sources: a2d2.audi https://lnkd.in/eUXSn_tU https://lnkd.in/eAVExMDG #AutonomousDriving #ADAS #DataIntelligence #SemanticSearch #UnderstandRawData #DataAcquisition #ComputerVision #ScenarioCatalogue #DataPreprocessing #DataExploration #Vectorisation #Vectorengine #Vectorisationengine #Vectorembeddings
-
Great summary of the AI Index report of the Stanford Institute for Human-Centered Artificial Intelligence. If you want to know what happened in 2023 in the AI space, have a look. https://lnkd.in/gK3T3Vue #GenAI #artificialintelligence #StateofAI #GenerativeAI #AI
AI Index: State of AI in 13 Charts
hai.stanford.edu