Visual Intention Grounding for Egocentric Assistants
This work addresses the challenge of enabling AI assistants to understand and ground implicit human intentions in egocentric views, which is incremental as it builds on existing multimodal LLMs with a new dataset and training approach.
The paper tackles the problem of visual grounding in egocentric AI assistants, where objects are referred to implicitly through intentions rather than explicit descriptions, by introducing the EgoIntention dataset and a Reason-to-Ground (RoG) instruction tuning method. The result shows that RoG significantly outperforms baseline methods on this new benchmark while maintaining performance on standard tasks.
Visual grounding associates textual descriptions with objects in an image. Conventional methods target third-person image inputs and named object queries. In applications such as AI assistants, the perspective shifts -- inputs are egocentric, and objects may be referred to implicitly through needs and intentions. To bridge this gap, we introduce EgoIntention, the first dataset for egocentric visual intention grounding. EgoIntention challenges multimodal LLMs to 1) understand and ignore unintended contextual objects and 2) reason about uncommon object functionalities. Benchmark results show that current models misidentify context objects and lack affordance understanding in egocentric views. We also propose Reason-to-Ground (RoG) instruction tuning; it enables hybrid training with normal descriptions and egocentric intentions with a chained intention reasoning and object grounding mechanism. RoG significantly outperforms naive finetuning and hybrid training on EgoIntention, while maintaining or slightly improving naive description grounding. This advancement enables unified visual grounding for egocentric and exocentric visual inputs while handling explicit object queries and implicit human intentions.