Computer Vision

Venaka develops applied computer vision systems that convert raw imagery — captured via smartphones, drones, fixed cameras, or field devices — into actionable, measurable insights for environmental and agricultural decision-making. Our work is grounded in data-centric AI, where model performance is the outcome of rigorous dataset design, robust evaluation, and deployment-aware engineering. We actively incorporate Vision Transformers (ViTs) and hybrid transformer-CNN approaches to address generalisation under domain shift, fine-grained categorisation, uncertainty-aware outputs, and efficient deployment.