Zedge's DataSeeds.AI Releases Foundational Dataset for Computer Vision and Generative AI in Collaboration with Perle.ai and Émet Research
- Release of high-quality AI training dataset with proven 70% performance improvement over benchmarks
- Access to massive catalog of 30M+ rights-cleared images through GuruShots and Zedge Premium
- Potential new revenue stream through B2B marketplace for AI datasets
- Superior annotation quality demonstrated by outperforming AWS Rekognition with detailed human expert reviews
- Scalable platform capable of launching on-demand photo challenges for custom dataset creation
- Early stage of commercialization with uncertain market adoption
- Competitive landscape in AI training data market could impact pricing and market share
- Dependency on GuruShots' photographer community for content generation
Insights
Zedge's strategic entry into AI training data market leverages existing content ecosystem for new B2B revenue streams.
Zedge's release of the DataSeeds.AI Sample Dataset (DSD) represents a strategic pivot into the high-value AI training data market. By monetizing content from their GuruShots photography platform, Zedge has created a commercially viable dataset product with demonstrable technical advantages over existing solutions.
The technical differentiation is substantial: the research paper indicates DSD-trained models achieved a 24.09% increase in BLEU-4 scores and 70% better results compared to benchmark datasets. AWS Rekognition's poor F1 score (0.19) against DSD annotations demonstrates the quality gap between automated tagging and DSD's human-expert approach.
What's particularly valuable is Zedge's scalable content pipeline. With access to over 30 million rights-cleared images and the ability to generate custom datasets through targeted GuruShots challenges, Zedge has positioned itself to respond rapidly to enterprise AI training needs. This creates a competitive advantage in the data-centric AI development paradigm where quality training data is increasingly recognized as the differentiating factor in model performance.
For Zedge, this represents a significant business model expansion beyond consumer-facing digital content. By transforming their existing UGC assets into enterprise-grade AI training resources, they've created a new revenue stream that leverages their core community assets while diversifying beyond their traditional marketplace model.
Dataset establishes new AI training benchmarks as detailed in accompanying research paper
NEW YORK, NY / ACCESS Newswire / June 9, 2025 / Zedge, Inc. (NYSE American:ZDGE), $ZDGE, a leader in digital marketplaces and interactive games that provide content, enable creativity, empower self-expression and facilitate community, today announced the release of a new foundational image dataset - DataSeeds.AI Sample Dataset (DSD) - purpose-built for computer vision and generative AI model training. The dataset was created in partnership with Perle.ai and Émet Research and represents a major step forward in data-centric AI image development.
Jonathan Reich, CEO of Zedge, Commented:
"The DSD release marks a significant milestone for DataSeeds.AI, whose goal is to become a major supplier to enterprises that create foundational models in need of rights-cleared, high-quality images. The DSD annotation delivers measurable improvements over legacy solutions like AWS Rekognition, setting a new benchmark for high-quality, human-aligned AI training data. DataSeeds.AI was able to assemble the DSD by leveraging GuruShots' tightly knit photographer community and their wide-ranging portfolio of photographs for high-quality AI training data. This release not only underscores the commercial potential of DataSeeds.AI as a serious contender in the evolving B2B marketplace arena for AI datasets but also highlights the natural synergies that exist with our creators across both GuruShots and the Zedge Premium marketplace. It validates our ability to turn user-generated content into scalable, enterprise-grade datasets that can generate new revenue sources for Zedge."
The DSD is comprised of over 7,800 high-quality photos sourced from players of Zedge's leading photography game, GuruShots. Every image in the dataset was ranked by players of the game, and subsequently, each image was annotated by expert reviewers who provided detailed descriptions of the image content. The DSD release marks a major step in building the kind of real-world, human-reviewed data that improves the veracity of modern AI models.
The introduction of the DSD highlights the inherent value in DataSeeds.AI's capacity to meet custom image demand promptly by launching relevant GuruShots photo challenges and/or by accessing existing images from GuruShots' massive catalog. Whether it is improving generative AI models, analyzing scenes or handling edge cases, the platform offers a scalable pipeline supported by tens of thousands of photographers that can provide diverse and rights-protected images.
Ahmed Rashad, CEO of Perle.ai remarked, "The DataSeeds.AI partnership allowed us to apply our methodologies, which leverage domain expertise and AI, for high quality data annotation and while validating the results through comprehensive benchmarking research. We are thankful for Zedge's partnership and the meaningful contribution that DataSeeds.AI is making to the AI community. DSD is a milestone for human-aligned dataset creation."
Freeman Lewin, CEO of Émet Research said, "We're deeply grateful to Zedge, DataSeeds.AI and Perle.ai for enabling this release. Together, we've not only demonstrated the power of data-centric AI but also introduced a best-in-class model for data to be used for AI training. We're excited to keep supporting important AI research efforts in conjunction with industry leaders like Zedge and Perle.ai."
The release of the DSD is accompanied by a evaluative research paper titled "Peer-Ranked Precision: Creating a Foundational Dataset for Fine-Tuning Vision Models from DataSeeds' Annotated Imagery," which shows how training AI models with the DSD yields
The DSD was labeled through a multi-tiered process where human experts described scenes in natural language and even outlined certain objects down to the pixel. This helps AI learn in a way that's closer to how people view and explain the world.
Technical Deep Dive: Research Findings and Differentiators
The DSD was designed to serve as a reproducible benchmark for training and fine-tuning multimodal vision-language models. It includes 7,843 high-resolution, rights-cleared photographs sourced from GuruShots, each selected through a unique in-game peer-ranking system that reflects aesthetic and compositional quality validated by a global photography community.
Each image was then enhanced with multi-tiered human annotation through Perle.ai's expert-in-the-loop pipeline, including:
Pixel-level segmentation
Structured scene descriptions
Technical metadata (e.g., exposure, focal length, depth-of-field assessments)
Title and category-level labels aligned to visual content
This combination of peer review, expert annotation and visual diversity enables DSD to provide context-rich training data that improves model grounding and multimodal comprehension.
Key empirical findings from the research paper include:
Fine-tuning LLAVA-NEXT on DSD led to a
24.09% increase in BLEU-4, with corresponding gains in ROUGE-L, BERTScore and CLIPScore, validating stronger semantic precision and image-text alignment.When benchmarked against the DSD annotations, AWS Rekognition achieved only a 0.19 F1 score, demonstrating the limitations of automated commercial tagging systems for high-quality dataset curation.
BLIP2 models also showed meaningful improvement when fine-tuned on the DSD, indicating that the dataset generalizes across different architectures and not just LLAVA-style models.
What makes the DSD and DataSeeds.AI uniquely valuable?
Data-centric development: The DSD supports the shift from model-centric to data-centric AI by prioritizing quality, context and diversity in training inputs.
Scalable generation: DataSeeds.AI can rapidly build domain-specific datasets by launching on-demand GuruShots photo challenges and/or by drawing from Zedge Premium's and GuruShots' massive catalog of 30M+ and growing, rights-cleared images enriched with EXIF metadata, tags and geolocation diversity.
Human-aligned annotation: Unlike auto-tagged datasets, the DSD annotations provide interpretability, nuance and grounding that support vision-language understanding and generalization to real-world use cases.
Open availability: All data, models and benchmarking results are reproducible and available on HuggingFace, encouraging adoption, validation and further innovation.
This foundation positions Zedge's DataSeeds.AI platform as a differentiated supplier of high-fidelity, human-reviewed datasets tailored to the evolving needs of the generative AI ecosystem.
Access the research paper here
Access the DSD here
About Zedge
Zedge empowers tens of millions of consumers and creators each month with its suite of interconnected platforms that enable creativity, self-expression and e-commerce and foster community through fun competitions. Zedge's ecosystem of product offerings includes the Zedge Marketplace, a freemium marketplace offering mobile phone wallpapers, video wallpapers, ringtones, notification sounds, and pAInt, a generative AI image maker; GuruShots, "The World's Greatest Photography Game," a skill-based photo challenge game; and Emojipedia, the #1 trusted source for 'all things emoji.' For more information, visit: investor.zedge.net
Follow us on X: @Zedge
Follow us on LinkedIn
About DataSeeds.AI
DataSeeds.AI offers both on-demand and off-the-shelf image and video datasets enriched with detailed metadata, perfectly suited for AI model training. By leveraging a vast global network of creators and an extensive catalog, we provide rapid data collection and diverse content, ensuring swift, scalable solutions that accelerate AI training. For more information, visit: https://www.DataSeeds.AI/
About Perle.ai:
Perle.ai provides expert data annotation and enrichment services for AI development. Leveraging a curated global network and AI-assisted workflows, Perle delivers high-quality, multimodal datasets built for real-world performance.
About Émet Research:
Émet Research provides sourcing, annotation, evaluation, research, compliance, licensing, liquidity, and sales solutions for data suppliers and AI labs around the world. Through its deep partnerships, and its own marketplace, Brickroad, Émet Research helps bring high-fidelity, proprietary datasets to market.
Contact:
Brian Siegel, IRC, MBA
Senior Managing Director
Hayden IR
(346) 396-8696
ir@zedge.net
SOURCE: Zedge, Inc.
View the original press release on ACCESS Newswire
FAQ
What is the DataSeeds.AI Sample Dataset (DSD) released by Zedge (ZDGE)?
How does ZDGE's DataSeeds.AI dataset improve AI model performance?
What is the size of Zedge's image catalog available for AI training?
How does DataSeeds.AI generate custom datasets for enterprise clients?
What makes ZDGE's DataSeeds.AI dataset unique in the market?
