meldCX Custom Computer Vision

A realistic 3D environment that simulates real world processes to train ML models with image datasets, and autonomously runs training cycles until it reaches baseline accuracy levels.

Advanced behavior detection capability can combine a predetermined set of actions and other data to detect anomalies and suspicious events, allowing teams to investigate and respond at the right time

Automation of synth & game engine simulation generates combinations of various scenarios that measure specific actions and appearance, with correlated data to train AI for behaviour detection capabilities.

Optimize production lines to scan, analyze, and manifest for mislabeling and defect detection. The feature is created for manufacturing vertical customers.

Automatically detect hazardous behavior in compliance with government safety regulations and alert management immediately to reduce risk of harmful incidents.

For products that are assembled manually, utilize a human assist camera to prevent errors. Each step completed can be verified and recorded to provide data for traceability along the journey.

To understand product and shelf engagement.

A coffee manufacturer realised they were sitting on a huge amount of loss per annum, due to human error in the product packaging stage. meldCX computer vision solution acts as a ‘human assist’ camera to detect and prevent these errors. Each step completed are verified and recorded on a dashboard to provide data for traceability. Use cases: product detection, product verification.

meldCX is working with a meat processing facility based in the USA, to use AI to deliver a manufacturing line quality check as meat cuts get packaged. meldCX's computer vision solution enables the facility to check that all meat products are accurately tagged and packaged, and that staff are adhering to the right processes in the packaging journey. Use cases: product detection, label recognition, product verification.

meldCX built a handwriting recognition solution for a postal service organisation, dubbed SALi (Smart Automated Lodgment API). SALi is trained using synthetic datasets in 3D, and can run automated training of its ML models using the synth, and learn to recognise real world parcels and manifests. Use cases: detect package, verifies package size & weight, reads handwritten information, detect the sender's identity, calculates shipping price.

https://www.youtube.com/watch?v=4H5y0aeCLWk&t=2s

A transport infrastructure company is working with meldCX to analyse and ensure that delivery trucks are leaving the truck yard with the accurate weight of materials e.g. asphalt. After the trucks are captured from license plate recognition (LPR) to baseline and net weight, the solution will prepare the data ready for transport through data logistics with the goal that the team can dispatch trucks accurately and efficiently for all ongoing infrastructure projects.