arXiv:2606.24225v1 Announce Type: new 
Abstract: Object-level geometric edits, including translating, rotating, scaling, duplicating, or removing an object, are routine operations in digital content creation (DCC) workflows, yet they remain unreliable in generative video editing. The key challenge lies in specifying the target object's 3D state change unambiguously across viewpoint and time, while consistently updating geometry-dependent secondary effects such as shadows and reflections. We introduce GIVE, a geometry-instructed video editing framework that represents edits through a unified object-state formulation. Two video-aligned geometry streams describe the target object before and after editing: a depth-box encoding coarse 3D placement and extent, and an orientation-box providing an appearance-agnostic orientation cue. Together, these streams provide a compact pre/post geometric specification for object-state transitions. To provide paired supervision for learning these edits, we build a scalable graphics-engine pipeline that executes object-level edit programs and renders controlled before/after pairs, isolating the intended geometric edit while keeping secondary effects consistent with the transformation. Experimental results demonstrate that GIVE produces faithful geometric edits with temporal coherence and consistent secondary effects across operators in a unified framework, and shows promising transfer to in-the-wild videos. Project page: https://geometry-instructed-video-editing.github.io/give/

تم تقديم إطار جديد يسمى GIVE لتحرير الفيديو المستند إلى الهندسة، مما يتيح إجراء تعديلات هندسية دقيقة على مستوى الكائن، مثل النقل والدوران والتكبير في تحرير الفيديو التوليدي. يتناول هذا الإطار التحدي المتمثل في تحديد تغييرات الحالة ثلاثية الأبعاد للكائنات المستهدفة بشكل غير غامض مع تحديث التأثيرات الثانوية المعتمدة على الهندسة مثل الظلال والانعكاسات.

Se ha introducido un nuevo marco llamado GIVE para la edición de video instruida por geometría, que permite ediciones geométricas precisas a nivel de objeto, como la traducción, rotación y escalado en la edición de video generativa. Este marco aborda el desafío de especificar de manera inequívoca los cambios de estado 3D de los objetos objetivo mientras actualiza los efectos dependientes de la geometría, como sombras y reflejos.

Un nouveau cadre nommé GIVE a été introduit pour l'édition vidéo instruite par la géométrie, permettant des modifications géométriques précises au niveau des objets, telles que la translation, la rotation et l'échelle dans l'édition vidéo générative. Ce cadre aborde le défi de spécifier de manière non ambiguë les changements d'état 3D des objets cibles tout en mettant à jour les effets dépendants de la géométrie comme les ombres et les réflexions.

A new framework named GIVE has been introduced for geometry-instructed video editing, enabling precise object-level geometric edits such as translation, rotation, and scaling in generative video editing. This framework addresses the challenge of unambiguously specifying 3D state changes of target objects while updating geometry-dependent effects like shadows and reflections.