Look-ahead before you leap: end-to-end active recognition by forecasting the effect of motion

TL;DR

This paper introduces an end-to-end active recognition system that learns to forecast the effects of its motions on its internal representation, improving recognition accuracy by actively choosing viewpoints.

Contribution

It proposes a recurrent neural network-based approach that jointly learns motion policies and forecasts view effects, advancing active recognition capabilities.

Findings

The system learns meaningful motion policies for recognition.

Forecasting view effects enhances recognition accuracy.

Results validated on two challenging datasets.

Abstract

Visual recognition systems mounted on autonomous moving agents face the challenge of unconstrained data, but simultaneously have the opportunity to improve their performance by moving to acquire new views of test data. In this work, we first show how a recurrent neural network-based system may be trained to perform end-to-end learning of motion policies suited for this "active recognition" setting. Further, we hypothesize that active vision requires an agent to have the capacity to reason about the effects of its motions on its view of the world. To verify this hypothesis, we attempt to induce this capacity in our active recognition pipeline, by simultaneously learning to forecast the effects of the agent's motions on its internal representation of the environment conditional on all past views. Results across two challenging datasets confirm both that our end-to-end system successfully learns meaningful policies for active category recognition, and that "learning to look ahead" further boosts recognition performance.