Showing 3 Result(s)

POC – Real-time deep learning segmentation

SiliconBased, a pioneering company based in Tarragona, embarked on a Proof of Concept (POC) that promises to transform the way hydrocarbon transport handles loading and unloading processes. At its core, this innovative project leverages the power of deep learning and computer vision to enhance safety and efficiency. A customized dataset was meticulously crafted to accommodate …

Foams, sludges and green algae monitoring

Monitoring wastewater treatment plants with AI has been one of the most successful projects from the innovation area of Veolia group. For this purpose we build custom datasets of segmentation of different problems and test instance and semantic segmentation architectures until we find the more efficient models in external conditions monitoring water 24/7. Lighting changes …

Waste recycling training

Updata.ca is a company from Québec that hired me for trying different architectures in their solution to classify waste recycling. I was in charge of testing instance segmentation models and they had a dataset of tens of millions images (artificially created). We trained models for days in 2 A100 (40Gb) cluster but results were not …