#Japan #Tokyo #AI Dataset #Qlean Dataset #Visual Bank

Visual Bank Introduces New Judo Video and Image Dataset

Visual Bank Inc., based in Shibuya, Tokyo, led by CEO Masayuki Nagai, is expanding its AI training data solutions through its subsidiary, Amana Images. The company has recently introduced a Judo Video and Image Dataset, further enriching its lineup of proprietary AI development data under the Qlean Dataset brand.

Overview of Qlean Dataset

The Qlean Dataset serves as a versatile resource for commercial use, tailored for various AI development needs. The recently added dataset includes high-quality video and image data capturing judo performances from multiple angles, enhancing developers' capabilities to work with sports-related AI models. The Evolving AI Data Recipe concept allows customizable datasets to suit different specifications, ensuring flexibility in usage and implementation.

Detailed Features of the New Dataset

The new Judo Video and Image Dataset features:

• - Content: Video and image captures of male judo athletes performing tasks from various angles.
• - Athlete Demographics: Covers Japanese male athletes in their 20s, including two middleweight and two heavyweight competitors, totaling four individuals.
• - Recording Environment: Captured indoors, within a dojo setting.
• - Action Coverage: Features a comprehensive array of techniques including 18 throwing techniques, 4 hip techniques, 4 sacrifice techniques, 9 ground techniques, 7 choke techniques, and 2 joint lock techniques.

Video Dataset Specifics

• - Data Size: 8.28GB
• - Record Count: 180 videos
• - File Type: MP4
• - Total Duration: 31 minutes and 57 seconds

The video data has a dedicated sample URL for users wanting to preview the contents.

Image Dataset Specifics

• - Data Size: 15.4GB
• - Record Count: 82 images
• - File Type: JPEG

For image previews, users can refer to the sample URL.

Use Cases for the New Judo Dataset

1. Pose Estimation and Skeleton Recognition Development: The intricacies of judo techniques make this dataset ideal for developing algorithms that recognize whole-body movements, essential for sports science and motion analysis AI.
2. Behavior Recognition and Interpersonal Motion Analysis: The dataset captures the dynamic flow of action in combat sports, providing rich data for training behavior recognition AI and enhancing security surveillance measures.
3. AI for Scoring and Judgement Assistance: With a complete catalogue of official judo techniques, this dataset allows automation in skill assessment and scoring, enhancing fairness and precision in competitive environments.
4. VR/AR and Metaverse Motion Generation: The angles and quality of recorded judo actions are conducive for creating realistic movement simulations in virtual or augmented reality settings, allowing immersive sports education experiences.
5. Robotics and Motion Imitation Learning: The dataset can support humanoid robotic research by training robots to mimic complex judo movements, thereby advancing the field of interactive melee robotics.

Features of Qlean Dataset

• - Supports Research and Commercial Use: All dataset contributors have provided consent for data access, ensuring compliance with privacy regulations globally, making the datasets safe for academic and commercial research.
• - Swift Data Procurement: Utilization of the unique AI Data Recipe framework means quicker access to customized datasets, optimizing return on investment while minimizing initial costs.
• - Custom Dataset Creation: For data not available within the existing dataset range, Qlean Dataset is capable of creating tailored datasets that meet specific research needs.

Academia Support Program

In addition, Visual Bank has initiated a free dataset provision program aimed at academia. Universities, research institutions, and non-profit development teams can access over 80 different data types, consisting of more than 500,000 pieces, solving the challenge of acquiring high-quality, rights-clear data for educational purposes. More information is available here.

About Visual Bank

Visual Bank is a startup committed to establishing a next-generation data infrastructure that maximizes AI development capability. With a mission to "unlock the potential of all data," the company also operates AI-assisted tools like THE PEN for manga artists, alongside its comprehensive dataset development services through Amana Images. Visual Bank has been selected for the national research and development program GENIAC, propelling its initiatives towards real-world applications.

CEO: Masayuki Nagai
Location: 5-3-23 Kojimachi, Chiyoda-ku, Tokyo, Japan
Visual Bank Website: visual-bank.co.jp
Amana Images Website: amanaimages.com