SECO Expands Edge AI Portfolio with New Applications for Safety, Energy-Efficient Vision, and Geospatial Analytics – EEJournal

Applications of AI


New applications extend SECO’s AI capabilities with solutions for environmental analytics, real-time safety, and efficient edge intelligence

Arezzo, Italy – December 2, 2025 – SECO has introduced a new set of edge AI applications that power safety monitoring, low-power machine vision, geospatial analysis, and intelligent video understanding. This release strengthens SECO’s growing ecosystem of production-ready AI solutions designed for deployment across industrial and embedded environments.

Newly added capabilities include Land Cover Change Detection, an advanced computer vision application that analyzes satellite images and aerial photography to automatically identify and map land cover changes such as deforestation and urbanization. This functionality is essential for environmental monitoring, sustainable urban planning, and ecological impact assessment.

SECO is also expanding into voice-driven workflows with Wave2Vec – Speech-to-Text, a high-precision transcription application based on the Wave2Vec model and optimized to convert speech to text. Supports use cases such as automated documentation in industrial and medical settings and voice-controlled operations in complex environments.

For energy-sensitive devices, Object Detection – Low Power at the Edge provides energy-efficient object detection capabilities designed for power-constrained edge platforms. It enables real-time recognition of objects, people, and events where battery life, portability, and thermal efficiency are important.

On the safety side, Human Fall Detection provides an AI vision application focused on detecting human falls in real-time using advanced models. This is particularly useful for elderly care and high-risk industrial environments, allowing for immediate response through automated alerts.

SECO is also expanding its video analytics capabilities with Video Action Recognition, a deep learning application that analyzes video sequences to recognize and classify complex human actions. Supporting workflow analysis, operational safety, and staff training from manufacturing to retail.

Rounding out the updates, Face ID – Training Experience provides a customizable training environment specifically for training and fine-tuning facial recognition (Face ID) models. This allows you to create, optimize, and validate identity verification solutions directly on the target hardware, ensuring accuracy and alignment with deployment requirements.

Each application is validated for deployment across platforms powered by Intel, Qualcomm, NXP, AMD, Rockchip, and MediaTek and is backed by SECO’s enterprise-grade long-term support to ensure stability, security, and smooth integration within SECO’s hardware ecosystem.

Fausto Di Segni, Head of IoT and AI at SECO, said: “Each new model addresses a specific need seen in the field, from energy efficiency to real-time safety, to help accelerate customer adoption.”

New AI applications can be found on the SECO Application Hub, complete documentation is available on the SECO Developer Center, and an opportunity to try Clea Framework for free is also provided.

Seco

SECO (IOT.MI) is a high-tech company that develops and manufactures cutting-edge solutions for the digitalization of industrial products and processes. SECO’s hardware and software products enable B2B companies to easily bring edge computing, Internet of Things, data analytics, and artificial intelligence to their business. SECO’s technology spans multiple application areas, serving more than 450 customers across sectors such as healthcare, industrial automation, fitness, vending, and transportation. Through live monitoring and smart control of field devices, SECO solutions contribute to greener business operations through more efficient use of resources. SECO provides comprehensive engineering, technical support and business services in the United States and Canada from its headquarters in Rockville, Maryland.





Source link