Visual Bank Launches Qlean Dataset with Japanese Castle Images for AI Research

Visual Bank Expands AI Data Solutions with New Dataset

Visual Bank Co., Ltd., based in Shibuya, Tokyo, has announced the launch of the 'Qlean Dataset', which is now enriched with a new addition: the 'Japanese Castle Exterior Image Dataset'. This initiative, which operates through its subsidiary Amana Images, aims to provide customizable AI training data solutions, thereby facilitating both research and commercial AI development.

Overview of the Qlean Dataset

The Qlean Dataset features original data that is suitable for commercial use. This comprehensive data lineup, termed 'Data Recipe', allows users to combine readily available data materials flexibly according to their specific needs—be it accuracy, deadlines, or custom specifications. The dataset not only includes annotated and non-annotated images but is also designed to cater to individual requirements through adjustments or expansions.

Visual Bank has formed partnerships with organizations like the Chiba Lotte Marines and Toyo Keizai, among others, to continuously enhance the offerings through domestic and international networks. This ensures that the burden of data collection and organization for AI development is significantly alleviated, thus contributing to accelerated innovation in the field.

Details of the Japanese Castle Exterior Image Dataset

Contents of the Dataset

- Data Type: Exterior images of castles located across Japan
- Image Resolution: Over 2,400 pixels in length
- File Format: JPEG
- Total Files: 2,000

Sample Image Access

For those interested in viewing a sample of this dataset, please visit Sample Image Link.

Potential Use Cases for the Dataset

1. AI Training for Architectural Style Classification
The clear features of castle images such as keep structures, stone walls, and turrets offer ideal training data for AI models focused on classifying historical architecture. This can support research institutions and local governments involved in cultural heritage management.

2. Digital Archiving of Cultural Resources
With a diverse collection of castles, this dataset can serve as a foundational resource for educational and museum digital archives, contributing to the preservation and display of cultural heritage.

3. Enhancement of Tourist AI Apps
The dataset can also be utilized as training data for automatic recognition features in tourist guide apps and AR navigation tools. It includes seasonal elements and structural characteristics, enabling adaptation to various tourist scenarios.

4. Base for Historical Architecture Generation AI
The structure of this dataset is perfect for the development of reproduction models of historical buildings in applications such as image generation, 3D restoration, and AR/VR. It encompasses seasonal landscapes featuring castles, stone walls, and cherry blossoms—important for learning about cultural beauty and historical context.

Benefits of Qlean Dataset for AI Development

- Compliance and Flexibility

The Qlean Dataset ensures that data is acquired with proper consent from all subjects, complying with privacy policies across different countries which makes it trustworthy for research and commercial applications.

- Speed and ROI Maximization

By utilizing the distinct 'Data Recipe' format, customers can expect a quick and cost-effective data procurement process, helping to optimize return on investment.

- Custom Data Solutions

For datasets not included in the 'Data Recipe', Visual Bank offers the capability to create and structure datasets according to specific individual requirements, leveraging their unique expertise.

Upcoming Event: Free Webinar on AI and Copyright Issues

Visual Bank will host a free online seminar on August 20, 2025, focusing on the legal risks of generative AI. Legal professionals from ZeLo and Qlean Dataset managers will discuss legal responsibilities involved in AI development, research, and implementation, alongside practical case studies. Those interested in learning more about data provision that respects rights are encouraged to register.

- Date: August 20, 2025
- Time: 14:00 - 15:30
- Format: Online (Free participation)
- Registration Link: Webinar Sign-Up

Visual Bank Co., Ltd. strives to build and provide a next-gen data infrastructure aimed at maximizing AI development capabilities. Their mission is to unleash the potential of all data, which includes diverse initiatives like 'THE PEN'—a project focused on IP×AI use cases—and the provision of the Qlean Dataset through Amana Images, which operates as a fully owned subsidiary. Furthermore, Visual Bank has been selected for the national research and development program 'GENIAC', significantly advancing its efforts toward societal implementation.

CEO: Masayuki Nagai
Location: 5-3-23 Kojimachi, Chiyoda-ku, Tokyo, WeWork, Nippon Television Yotsuya Building
Company URL: Visual Bank
Amana Images URL: Amana Images