Qu’est-ce qu’Adobe Sensei ?

Adobe Sensei peut exploiter, grâce à l’intelligence artificielle, des milliers de milliards de données et de contenus — depuis les images haute résolution jusqu’aux clics des clients — au sein d’un framework unifié d’intelligence artificielle et d’apprentissage automatique. Rechercher des correspondances entre des millions d’images, comprendre la signification et l’orientation d’un document ou cibler précisément des segments d’audience importants…



Adobe Sensei sur l’ensemble de la plate-forme cloud d’Adobe

Dans Adobe Creative Cloud, Adobe Sensei anticipe l’étape suivante. Il recrée sur les photos des éléments qui n’y figuraient pas, en analysant les pixels voisins. Il reconnaît la typographie et recrée des polices. Il identifie des objets sur vos images et ajoute des termes indexables aux balises des photos. De plus, il reconnaît les visages et place des repères sur les sourcils et les lèvres, dont l’expression peut ensuite être modifiée d’un simple clic. Dans Adobe Document Cloud, il transforme les documents papier en fichiers numériques modifiables. Il inclut automatiquement les bonnes polices, crée des champs de formulaire et « nettoie » les signatures.

Adobe Sensei and Adobe Cloud Platform

Adobe Sensei is a framework and set of intelligent services built into the Adobe Cloud Platform which dramatically improve the design and delivery of digital experiences. Adobe Sensei is already powering Adobe products such as Photoshop and our Digital Marketing Cloud and will be available as APIs via Adobe I/O.



Partners and developers will be able to build entirely new types of applications and solutions for their customers by taking advantage of Adobe’s history of innovation around AI and machine learning. Examples of intelligent APIs include :

  • Content intelligence: Uses deep learning to search and tag images automatically, and makes intelligent recommendations when a user searches for images
  • Face Aware Editing: Finds faces in an image and uses “landmarks” such as eyebrows, lips and eyes to understand their position and change the facial expression without ruining the image
  • Semantic Segmentation: Shows each image region labeled with its type; for example, whether it is a building or the sky. Such labelled regions allow easy selection and manipulation of objects, using simple commands (e.g., “change the sky”)
  • Natural Language Processing: Provides text understanding, topic modeling and sentiment analysis of digital documents
  • Document Similarity: Recognizes similar documents and highlights their differences
  • Optical Character Recognition: Converts and cleans up documents that have been scanned from paper and removes ”dust” artifacts, or automatically straightens crooked lines of text.
  • Attribution: Algorithmically determining the impact of different marketing touch points on consumers’ decisions to engage with a brand.
  • Anomaly Detection: Simplifies data analysis by surfacing the most relevant insights and highlighting anomalies using statistics to focus on what needs attention
  • Sentiment Analysis: Helps see and predict what customers like, talk about, and share most.