Zedge, through its AI data division DataSeeds.AI, announced on June 9, 2025, the release of a new large-scale, foundational dataset. It is designed for training and testing computer vision and generative AI models. According to a press release distributed via Access Newswire and noted by financial portal StockTitan, this project was carried out in close collaboration with research partners Perle.ai and Émet Research. High-quality, diverse, and meticulously annotated datasets are the foundation for creating powerful and accurate AI models, and their scarcity often becomes a bottleneck in research. The new dataset from DataSeeds.AI aims to solve this problem by providing the developer and researcher community with well-labeled visual data necessary for training the next generation of AI systems. The set reportedly contains a large number of unique images with rich, multi-layered annotations that go beyond simple labels and may include detailed object segmentation, scene descriptions, and semantic relationships between objects. The partnership with Perle.ai and Émet Research likely ensured a high level of scientific and technical expertise in data collection, processing, and annotation, as well as quality control. It is expected that the release of this dataset will significantly improve the performance of existing computer vision models, reduce their bias, and enhance the quality, realism, and controllability of generative AI. This release is a significant contribution to the development of the entire AI ecosystem by providing developers with higher-quality "fuel" for their models.
Zedge (DataSeeds.AI) Releases New Foundational Dataset for AI
