AI+ Tool and Solution Catalogue

Reference No. A-0193
Solution Name No Deep Learning Predictive Maintenance

Solution Description

The Mean Time Between Failures (MTBF) for modern machinery can be very high. It can take years for equipment parts to wear out a collapse.It is therefore virtually impossible to collect sufficient datasets of deteriorating equipment to build AI models for predicting when they require maintenance.

WildFaces’ Predictive Maintenance system requires very small datasets for building predictive models for estimating when equipment will require maintenance.More importantly we can provide Preventive Maintenance Systems as we can detect anomalies in the look, sound and smell of equipment and can use that to predict potential deterioration of performance.Our Predictive Maintenance AI, based on multi-sensory AI utilizing video, sound and smell analytics, can automate labour-intensive inspection processes.

WildFaces’ unique “Intuitive AI” approach requires NO Deep Learning, instead emulating complex human intelligence to address real-life complexity such as in equipment maintenance arena. As equipment failures happen at rare instances it is not possible to capture any defective data. As such, the deep learning approach is NOT practical as the starting point is usually the collection of massive datasets about the defective parts of each particular component.

WildFaces’ No Deep Learning “intuitive AI” can ensure successful AI adoption in the Maintenance space as it requires less than 30 images of components in good condition for training each component model. There is also no requirement of expensive and power-hungry GPUs making AI deployment portable and cost-effective.Our No-Deep-Learning Anomaly detection methodology detects wear & tear from normal operating performance using video, sound and smell analytics.

The system looks for dents, rust, scratches, loosening parts, thickening/thinning parts, changing color, oil leakage/ water leakage, vapor/smoke etc. Our Sound AI can detect 100+ of non-conversational maintenance related sounds such as vibration, loud engine, backfire, glass breaking sound and etc. For smell AI, we can pick up any complex chemicals that could be of a concern such as toxic chemicals.

WildFaces’ patented “On-The-Move” WildAI capabilities works on IoT devices that can be mounted alongside Pan-Tilt-Zoom cameras, drones, moving robots and body-worn cameras.

Application Areas
Data Analysis and Prediction

Use Case

• Building Maintenance: Equipment involved can be escalators, elevators, HVAC etc. These will include Facility Management companies, service companies as well as the equipment manufacturers • Transportation Industry: Road Signs, Traffic Control & Signaling equipment, and Railway related equipment that has a short repair and maintenance time • Utility Maintenance: Equipment used throughout their utility plants • Construction Industry: Equipment used during Construction Phase • Manufacturing plants with heavy machinery • Plus more

Support On-premise deployment Yes
Support standalone laptop Yes
Graphics Processing Unit (GPU) required No
Pricing Model
Subscription-Based Pricing – Monthly/Annual Subscription
Free Trial Yes
Company / Organisation Name WildFaces Technology Limited
Email Address bd@wildfaces.ai
Telephone Number +85282162219
Website www.wildfaces.ai
Address Unit 608-613, Level 6, Core C, Cyberport 3, 100 Cyberport Road, Pokfulam, Hong Kong
Presentation Decks
Supplementary Files

Presentation Videos

If any government department would like to obtain additional information about the AI solution, please contact Smart LAB.