SSCNav: Confidence-Aware Semantic Scene Completion for Visual Semantic Navigation
This work tackles the problem of efficient semantic navigation for autonomous agents in unknown indoor environments, offering an incremental improvement over existing methods.
This paper addresses visual semantic navigation, where an agent must navigate to a target object category in an unknown environment. The proposed SSCNav algorithm uses a confidence-aware semantic scene completion module to infer a complete scene representation and guide navigation, which improves the efficiency of downstream navigation policies.
This paper focuses on visual semantic navigation, the task of producing actions for an active agent to navigate to a specified target object category in an unknown environment. To complete this task, the algorithm should simultaneously locate and navigate to an instance of the category. In comparison to the traditional point goal navigation, this task requires the agent to have a stronger contextual prior to indoor environments. We introduce SSCNav, an algorithm that explicitly models scene priors using a confidence-aware semantic scene completion module to complete the scene and guide the agent's navigation planning. Given a partial observation of the environment, SSCNav first infers a complete scene representation with semantic labels for the unobserved scene together with a confidence map associated with its own prediction. Then, a policy network infers the action from the scene completion result and confidence map. Our experiments demonstrate that the proposed scene completion module improves the efficiency of the downstream navigation policies. Video, code, and data: https://sscnav.cs.columbia.edu/