KINEVA – REBOTNIX

What is KINEVA?

KINEVA is the REBOTNIX platform that solves various industrial challenges with high-precision automation tasks and state of the art neuronal networks. The trained KINEVA models are released as opensource.

It offers powerful software and on-premise no-code tools. The HTML5 based application contains project and user management, image and data annotation, training engine, model analyse and export model components, all in one package.

How KINEVA works?

KINEVA enables users to annotate data and create AI automation without programming skills while ensuring the highest standards of data security.

The platform can include generative local models to manage data pipelines, securely transferring tasks from digital screens to physical machine control, enhancing efficiency and productivity.

KINEVA plus GUSTAV edge hardware is our 360 degree solution

The following video provides a 120-second overview of how KINEVA operates. First video on the left shows our KINEVA trailer, the video on the right shows how our pre-trained model performs.

KINEVA and UNREAL 5.5 engine

𝐔𝐧𝐫𝐞𝐚𝐥 𝐄𝐧𝐠𝐢𝐧𝐞 𝟓.𝟓 𝐚𝐧𝐝 𝐊𝐈𝐍𝐄𝐕𝐀: 𝐀 𝐍𝐞𝐰 𝐒𝐭𝐚𝐧𝐝𝐚𝐫𝐝 𝐟𝐨𝐫 𝐀𝐈 𝐓𝐫𝐚𝐢𝐧𝐢𝐧𝐠

With the release of Unreal Engine 5.5, we’re achieving a major milestone in AI training. The quality of synthetic scenes now enables a direct integration between our KINEVA Enterprise Training Software and Unreal Engine.

𝐄𝐟𝐟𝐢𝐜𝐢𝐞𝐧𝐭 𝐀𝐮𝐭𝐨-𝐥𝐚𝐛𝐞𝐥𝐢𝐧𝐠: Scenes created in Unreal Engine can be processed directly – without intermediate steps like saving individual JPG files.

𝐕𝐞𝐜𝐭𝐨𝐫 𝐃𝐚𝐭𝐚𝐛𝐚𝐬𝐞: Annotations and image data are combined in real-time, saving time and resources.

𝐂𝐨𝐦𝐩𝐫𝐞𝐡𝐞𝐧𝐬𝐢𝐯𝐞 𝐃𝐚𝐭𝐚 𝐏𝐫𝐨𝐜𝐞𝐬𝐬𝐢𝐧𝐠: KINEVA handles not only image data processing but also project management, training, and test inferencing – all in one platform.

𝐅𝐥𝐞𝐱𝐢𝐛𝐥𝐞 𝐈𝐧𝐟𝐫𝐚𝐬𝐭𝐫𝐮𝐜𝐭𝐮𝐫𝐞: Although KINEVA runs entirely in the browser and supports cloud-based setups, the primary focus is on in-house deployment on private GPU servers. This ensures that model ideas, training data, annotations, and the resulting models remain under the customer’s ownership.

𝐀𝐩𝐩𝐥𝐢𝐜𝐚𝐭𝐢𝐨𝐧𝐬:
From safety-critical areas, such as training person-related datasets with synthetic models, to traffic signs, potholes, metal surfaces, and robotic actions: KINEVA enables companies to create and manage model datasets for a wide range of scenarios.

More and more companies are recognizing the value of local applications tailored to their business models. But data alone is not enough – the key lies in transforming it intelligently into value-generating business solutions.

KINEVA workflow

We love opensource

We release some KINEVA per-trained AI as opensource.

The github link is located here.

GUSTAV hardware edge series for KINEVA

REBOTNIX, we go beyond software. We enhance KINEVA with our GUSTAV Edge (link) series, based on NVIDIA Jetson. The new GUSTAV MINI hardware is coming this August. It hosts certified cameras, sensors, and position locators with top IP protection. The inference engine of GUSTAV supports exported KINEVA models out of the box.

Cases

eFD bottle controlling uses a KINEVA use case where they develop an AI visual computing model to control the color and quality of the bottle cap with realtime reporting and sorting.

Quantum is a KINEVA-based AI use case for copper production control, designed for communication of visually detected events between robotics and their coordinate system. It is an autonomous AI model that coordinates and sorts copper plates during a 24/7 production scenario.

This system ensures continuous, efficient operations by managing the sorting process, maintaining seamless communication, and providing long-term storage of all relevant production data.

REBOTNIX Smart City Solutions is an application that is currently installed in 60 cities in Germany, Spain and the USA.

The AI model was trained from scratch with KINEVA and includes object recognition, segmentation and other classes.

Want to test KINEVA trainings plattform?

Please to fill our the form to get access to KINEVA demo session.