Towards Open World Recognition

TL;DR

This paper introduces the open world recognition problem, formalizes its challenges, and proposes the NNO algorithm that incrementally learns new categories while detecting outliers, validated on large-scale datasets.

Contribution

It formally defines open world recognition, extends existing algorithms, and introduces the NNO method for incremental learning and outlier detection in dynamic environments.

Findings

NNO outperforms existing methods on large-scale datasets.

Theoretical framework balances open space risk and empirical risk.

Validated on ImageNet with 1.2 million images.

Abstract

With the of advent rich classification models and high computational power visual recognition systems have found many operational applications. Recognition in the real world poses multiple challenges that are not apparent in controlled lab environments. The datasets are dynamic and novel categories must be continuously detected and then added. At prediction time, a trained system has to deal with myriad unseen categories. Operational systems require minimum down time, even to learn. To handle these operational issues, we present the problem of Open World recognition and formally define it. We prove that thresholding sums of monotonically decreasing functions of distances in linearly transformed feature space can balance "open space risk" and empirical risk. Our theory extends existing algorithms for open world recognition. We present a protocol for evaluation of open world recognition systems. We present the Nearest Non-Outlier (NNO) algorithm which evolves model efficiently, adding object categories incrementally while detecting outliers and managing open space risk. We perform experiments on the ImageNet dataset with 1.2M+ images to validate the effectiveness of our method on large scale visual recognition tasks. NNO consistently yields superior results on open world recognition.