Computer Vision

Vision is maybe the most pervasive sense in human life and, along with visual system, it carries out a lot of different and complex tasks that go from motion detection to pattern recognition and higher level scene interpretation. Processing of visual inputs allows human brain to collect information from the surrounding that it can use to build knowledge of reality and decision mechanisms.

Computer Vision encompasses all those techniques that try to reproduce mechanisms and outcomes of human visual system, implementing algorithms that enable machines to gain a more complex level of understanding of digital images and videos, producing quantifiable outputs. In the research field of motion analysis for cognitive and behavioral studies a video-based approach may be winning over sensor-based approach being less invasive and preserving the spontaneity of movement itself.

Applying this approach, Ab.Acus develops tailor made solutions as digital support for research in this field. Pattern recognition, object detection, pose estimation and tracking techniques have been used to support the motion analysis of adults, infants, fetuses, animals and plants.


Movidea: A Software Package for Automatic Video Analysis of Movements in Infants at Risk for Neurodevelopmental Disorders


TrHandy is a solution developed for a research project in the field of Reaching and Grasping study with the goal of studing the evolution of reach and grasp movements in children.


SPROUTS is a custom software developed inside a wide research project in collaboration with academic that aims at studying plants behavior.


Videotrack is a software package for the motion analysis of persons, animals or objects in any videoclip. It allows to cut time selections of the whole videoclip and examine the movement of interesting landmarks.