Komori
Other Disclosures3652Digital Media Professionals Inc.Information and Communications, Services and Others

DMP Conducts Demonstration Testing of an LLM Anomalous Behavior Detection System at GRAND GREEN OSAKA

DMP conducted a demonstration test of an LLM-based skateboard riding detection system at Grand Green Osaka with Nishio Rent All Co., Ltd. and others. Detection accuracy improved from 41.8% to 83.5%, while the number of detected skateboard riding incidents fell to 0 in April following the introduction of warning audio.

Worth reading
58In focus
0–100. This measures how worthwhile something is to read, not how good or bad the company is.
Direction
Improvement
Direction inferred from the company’s description.
Scale
83.5
Detection accuracy (%, as of the end of April 2026)
Wording
Definitive
How definitive the wording is.

Announced details

Vision-LLM InsightAI skateboard detection system
Applicable field (uses covered by this announcement)
Urban parksCommercial facilitiesConstruction sitesPublic spacesRoad and facility managementLogistics operations
Launch dateSeptember 2025 (full-scale provision of Vision-LLM Insight, actual results) / October 2025 (rental provision of the AI skateboard detection system, actual results)
What the company reported numerically
Quantitative Effects

Detection accuracy: improved from 41.8% to 83.5% (as of the end of April 2026)

Number of detected skateboard riding incidents: decreased from 71 (February) to 55 (March) and 0 (April)

Number of false positives at night: decreased from 20 over 63 days to 3 over 22 days

Relationship between the counterparty and the company
Digital Media Professionals Inc.3652
Relationship with the Company

DMP and 西尾レントオール株式会社 are business partners. 西尾レントオール株式会社 is responsible for conducting the demonstration and operating it on site, while DMP is responsible for developing and tuning the detection system. 西尾レントオール株式会社 provides the AI skateboard detection system for rental through DMP. General Incorporated Association Umekita MMO provided support in coordinating the implementation.

Counterparty西尾レントオール株式会社General Incorporated Association Umekita MMOUmekita 2nd District and Other Smart City Development Council

西尾レントオール株式会社 is headquartered in Osaka, Osaka Prefecture, and its representative director and president is Nishio Koji (西尾 公志). General Incorporated Association Umekita MMO is the area and park management organization for Grand Green Osaka. The Umekita 2nd District and Other Smart City Development Council provides overall coordination for the Smart City Implementation Support Project. Other attributes are Not disclosed.

Location (as indicated in this announcement)

Part of Grand Green Osaka (Kita-ku, Osaka)

Impact on Business Performance

The impact on business performance is Not disclosed

The company does not address Contract Value / Offering price.

Announced details
ItemDetails
Names of Products / Services CoveredVision-LLM Insight, AI skateboard detection system
Counterparty西尾レントオール株式会社, General Incorporated Association Umekita MMO, Umekita 2nd District and Other Smart City Development Council
Counterparty overview西尾レントオール株式会社 is headquartered in Osaka, Osaka Prefecture, and its representative director and president is Nishio Koji (西尾 公志). General Incorporated Association Umekita MMO is the area and park management organization for Grand Green Osaka. The Umekita 2nd District and Other Smart City Development Council provides overall coordination for the Smart City Implementation Support Project. Other attributes are Not disclosed.
Relationship with the CompanyDMP and 西尾レントオール株式会社 are business partners. 西尾レントオール株式会社 is responsible for conducting the demonstration and operating it on site, while DMP is responsible for developing and tuning the detection system. 西尾レントオール株式会社 provides the AI skateboard detection system for rental through DMP. General Incorporated Association Umekita MMO provided support in coordinating the implementation.
Launch dateSeptember 2025 (full-scale provision of Vision-LLM Insight, actual results) / October 2025 (rental provision of the AI skateboard detection system, actual results)
Applicable areaUrban parks, Commercial facilities, Construction sites, Public spaces, Road and facility management, Logistics operations
LocationPart of Grand Green Osaka (Kita-ku, Osaka)
Quantitative EffectsDetection accuracy: improved from 41.8% to 83.5% (as of the end of April 2026), Number of detected skateboard riding incidents: decreased from 71 (February) to 55 (March) and 0 (April), Number of false positives at night: decreased from 20 over 63 days to 3 over 22 days
Impact on Business PerformanceThe impact on business performance is Not disclosed
Items this document does not state (9)
Contract ValueOffering priceContract DateAward/certification nameAwarding / certifying bodyRegistration number and expiry dateLocation scaleUsage historySource

These are items the company does not state in this document. We do not leave them blank, nor fill them in from other sources or by inference.

The reasons the company gives

The company's own explanation, not something inferred from the numbers.

The detection logic was improved multiple times based on false-positive patterns collected in an actual urban environment, and a two-stage assessment process using generative AI was introduced. The combination with the warning audio function also confirmed a deterrent effect against nuisance behavior.

How to read this page

This is not the documentKomori reads the disclosure the company filed and writes this summary. Figures are as stated by the company and are not audited.
Weight is not a verdictIt measures how much this changes what the company had already told the market. It is not a view on the shares.
Corrections happenIf the company reissues this disclosure, the page is rebuilt from the corrected version and the revision number moves up.
DMP Conducts Demonstration Testing of an LLM Anomalous Behavior Detection System at GRAND GREEN OSAKA | Digital Media Professionals Inc. (3652) timely disclosure | Komori