Visual Bank Launches Snowboarding AI Training Dataset to Enhance Research and Development

Visual Bank Unveils New Snowboarding Dataset for AI Development

Visual Bank Inc., located in Minato, Tokyo, is pleased to announce the addition of a new dataset focused on snowboarding, specifically the halfpipe discipline, as part of its Qlean Dataset initiative. Managed under CEO Shinji Nagai, this offering aims to support both research and commercial AI development by providing high-quality training materials.

Overview of the Snowboarding Dataset

The newly launched Snowboarding (Halfpipe) Video and Image Dataset features content captured outdoors in a ski resort, focusing on two Japanese male snowboarders in their twenties.

Dataset Specifications

- Video Content: 28 clips totaling approximately 16 minutes and 24 seconds, provided in MP4 format with a total data size of 3.17 GB.
- Image Content: A collection of 8,683 high-resolution JPEG images, with a total data size of 25.3 GB.

For further details, check the sample URLs: Video Dataset and Image Dataset.

Use Cases for the Snowboarding Dataset

This dataset presents numerous opportunities for AI applications in various fields. Here are a few key use cases:

1. Aerial Pose Estimation and Trick Analysis

The dataset captures aerial rotations and posture changes from two angles, making it ideal for developing AI that analyzes body axis, center of gravity, and posture control during tricks. Sports science professionals can utilize this data for optimizing aerial forms and evaluating posture stability.

2. Competition Scoring and Trick Evaluation AI

By automating the extraction of scoring elements such as jump angles, rotation count, and posture maintenance duration, researchers can enhance the development of AI models that assist with scoring in competitions. These systems can provide valuable analytical support to training facilities and athlete performance assessments.

3. Motion Tracking and High-Speed Analysis AI

The high-speed motion capture, including reflections on snow, makes it suitable for verifying tracking accuracy and object recognition. This dataset can be utilized by drone and camera manufacturers for research into object-tracking algorithms.

4. VR/AR and Sports Video Generation AI Training

By combining multi-angle footage with high-resolution stills, developers can create realistic simulations of snowboarding in VR/AR environments. The dataset serves as an excellent training resource for generative AI involved in video synthesis and can be used in the development of competition experience content and sports education simulations.

5. Robotics and Motion Imitation Learning Research

The information concerning posture control during jumps and landing can significantly benefit robotics, specifically in balance control for robots and humanoids. The dataset can also serve as validation data for algorithms related to stabilization and motion control in aerial maneuvers.

Features of Qlean Dataset

Visual Bank's Qlean Dataset distinguishes itself through several key features:

- Research and Commercial Use: All datasets come with usage agreements compliant with privacy policies, ensuring safe research and commercial applications.
- Fast and Cost-Effective Data Procurement: With the unique AI Data Recipe, users can obtain datasets tailored for specific needs without heavy initial investments.
- Custom Dataset Creation: If a specific dataset is not available, Qlean can create optimized datasets based on individual requirements.

For assistance and inquiries about Qlean Dataset, visit Qlean Dataset Contact Form.

Qlean Dataset Academia Support Program

Visual Bank is also launching a complimentary dataset program as part of its commitment to academia. Targeting universities, research institutions, and non-profit tech development teams, this initiative offers over 500,000 data points across various formats like images, audio, video, and text. It aims to address the prevalent issues of acquiring high-quality, rights-cleared training data for research environments. Learn more about this initiative here.

About Visual Bank

Founded as a startup dedicated to building next-generation data infrastructure to maximize AI development capabilities, Visual Bank aims to unlock the potential of all data. In addition to the Qlean Dataset, which is primarily developed by its wholly owned subsidiary, Amana Images, Visual Bank provides AI-assisted tools for manga artists, helping them realize their creative visions. The company is also a participant in the national GENIAC research and development program, aiming to fast-track its contributions to society.