MemCam: Memory-Augmented Camera Control for Consistent Video Generation

2026-03-27

AI Summary

This work addresses the degradation of scene consistency in interactive video generation under dynamic camera control, particularly over long durations. To mitigate this issue, the authors propose a memory-augmented architecture that treats previously generated frames as external memory and employs context compression together with a co-visibility–based historical frame selection mechanism. This design enables efficient retrieval of long-range, relevant, and compact contextual information to guide camera motion. The approach maintains computational efficiency while significantly improving scene consistency in long videos and under large-angle camera movements, outperforming existing baselines and open-source state-of-the-art methods on key evaluation metrics.

Abstract

Interactive video generation has significant potential for scene simulation and video creation. However, existing methods often struggle with maintaining scene consistency during long video generation under dynamic camera control due to limited contextual information. To address this challenge, we propose MemCam, a memory-augmented interactive video generation approach that treats previously generated frames as external memory and leverages them as contextual conditioning to achieve controllable camera viewpoints with high scene consistency. To enable longer and more relevant context, we design a context compression module that encodes memory frames into compact representations and employs co-visibility-based selection to dynamically retrieve the most relevant historical frames, thereby reducing computational overhead while enriching contextual information. Experiments on interactive video generation tasks show that MemCam significantly outperforms existing baseline methods as well as open-source state-of-the-art approaches in terms of scene consistency, particularly in long video scenarios with large camera rotations.

Problem

• video generation
• scene consistency
• camera control
• contextual information
• interactive video

Innovation

• memory-augmented video generation
• camera control
• scene consistency
• context compression
• co-visibility-based selection