Visual Bank Co., Ltd. (Shibuya-ku, Tokyo; CEO Nagai Masayuki; hereinafter referred to as “Visual Bank”) is promoting the provision of “Qlean Dataset*,” an AI learning data solution that can be used for all kinds of research and commercial AI development through its subsidiary Amana Images Co., Ltd.
We have recently added the “Basketball Game Video Data Set” to our lineup, and are expanding our uniquely constructed AI development data lineup, “AI Data Recipes*.”
*Qlean Dataset:https://qleandataset.visual-bank.co.jp/
*AI Data Recipe:https://qleandataset.visual-bank.co.jp/lineup
About “AI Data Recipes” in “Qlean Dateset”
“AI Data Recipe” is an original data lineup that can be used commercially in “Qlean Dataset.”
It features a configuration that allows you to flexibly combine ready-to-use data materials depending on the purpose, accuracy and delivery date, and can also be used for partial annotated/unauthored data, as well as configuration changes and expansions according to individual requirements.
We are also expanding our lineup through partnerships with Chiba Lotte Marines Co., Ltd. and Toyo Keizai Shinpo Co., Ltd., domestic and international networks, and new recordings.
This greatly reduces the burden on data collection and maintenance at AI development sites, and contributes to accelerated development.
Overview of the “Basketball Game Video Data Set” that will be available this time
-
Subject attribute: 5 men in their 20s to 40s
-
Data capacity: 25.228GB
-
Number of data: 28
-
Data format: mp4
-
remarks
[撮影時間] 2 hours 5 minutes 36 seconds
[撮影環境] Gymnasium basketball court
[撮影距離・角度] Four fixed points on the basketball court: A/B/C/D
-
Sample details URL:https://qleandataset.visual-bank.co.jp/lineup/ds-010
-
Sample image:
“Basketball game video data set”Use case image of
-
Tactical analysis and formation recognition
It can be used as AI training data that uses match footage recorded from multiple perspectives to automatically analyze changes in the movements and formation of the entire team.
-
Comparative study of 3×3 and 5×5
Since it includes both three- and five-person systems, it is useful for AI research that analyzes the differences in play speed and space utilization. You can also learn about the formation density and differences in distance between athletes.
-
Detection and classification of player behavior
It can be used as the foundation data for AI models that accurately recognize unique actions such as dribbling, shooting, rebound, and defense.
-
A general-purpose study of behavioral recognition AI
Using video where multiple people move at the same time, we can develop behavioral recognition models that can be used in fields other than sports. It can also be applied to human flow analysis and multi-agent simulation.
-
Automatic editing during broadcasting and streaming
It can be used to develop automatic editing AI, such as highlight extraction and camera switching optimization, making use of fixed-point camera footage.
-
Referee support/fair play judgment
It can be used for research and development of AI models that detect fouls and goal decisions. It contributes to improving consistency in judgments and preventing mistakes in judgment.
-
VR/AR sports experience expansion
Based on the recorded footage, it is possible to develop VR training materials and AR viewing apps. It can be used to enhance your viewing experience by immersive experiences and real-time information display.
-
Camera technology development (autofocus tracking)
It uses high-speed movements within the gymnasium and simultaneous movement of multiple athletes, helping to improve the accuracy of autofocus and subject tracking algorithms. It can also be applied to sports cameras and surveillance camera development.
Features of dataset provided by Qlean Dataset
-
Supports research and development and commercial use
The dataset provided by Qlean Dataset has obtained consent forms for data acquisition and use for AI development from “all subjects”, and it also complies with the privacy policies of each country, so you can use it for research and commercial use with confidence.
-
Provides datasets from “AI Data Recipes” to speed up and maximize ROI
By taking the unique provision form of Qlean Dataset, called AI Data Recipes, you can procure data with a low initial investment.
-
Data sets not in the AI Data Recipes lineup can be created and constructed according to individual requirements.
For highly unique data, we can utilize the capabilities of Qlean Dataset to provide individually optimized datasets with requirements.
Qlean Dataset Contact Form:https://qleandataset.visual-bank.co.jp/contact
Qlean Dataset Service Site URL:https://qleandataset.visual-bank.co.jp/
Qlean Dataset Academy Support Program
Visual Bank has launched a free dataset offering program for its AI learning data solution, “Qlean Dataset,” as part of its academia support.
The target is universities, research institutes, non-profit technology development teams, and provides over 500,000 data items, including images, audio, video, and text.
As a company that has adopted GENIAC, Qlean Dataset will help resolve the issues that “we cannot obtain high-quality, clear-cut learning data” in research fields.
For more information, click here:https://qleandataset.visual-bank.co.jp/academia
Visual Bank Inc.
As a startup company that builds and provides next-generation data infrastructure that maximizes AI development capabilities, it operates its business activities with the mission of “unlocking all the possibilities of data.” In addition to “THE PEN,” which provides AI auxiliary tools to support manga artists who want to draw more, it also owns Amana Images Co., Ltd., which provides the AI learning data set development service “Qlean Dataset.”
Visual Bank was also adopted as a national research and development program, “GENIAC,” and is accelerating efforts to implement society.
CEO: Nagai Masayuki
Address: Nippon TV Yotsuya Building, 5-3-23 Kojimachi, Chiyoda-ku, Tokyo 102-0083 WeWork
Visual Bank Enterprise URL:https://visual-bank.co.jp/
Amana Images Company URL:https://amanaimages.com/about/
