Ekphrasis

2025, 6 min., 4K.

With the support of the Filmoteca de Catalunya.

With this installation, we continue a line of research that explores new algorithmic connections between words and images established by computer vision models. In Ekphrasis, fragments from various films such as Agnès Varda’s Documenteur, Jacques Tourneur’s Cat People and others are subjected to a complex and multi-layered algorithmic description process: objects, faces and emotions are detected and classified by object recognition systems, while image-to-text models are used to generate detailed descriptions. The piece draws attention to the processes of data mining in the digital world and to the poetic potential inherent in the radical impossibility of translating images into words and words into images.