# BAOD: Budget-Aware Object Detection

## TL;DR

This paper introduces BAOD, a budget-aware object detection framework that optimally selects images and annotations to train detectors efficiently under annotation constraints, achieving near-supervised performance with less labeling effort.

## Contribution

It proposes a hybrid supervised learning approach combined with an optimization-based data selection method for budget-aware object detection, outperforming other sampling techniques.

## Key findings

- Optimization-based selection outperforms random and active learning methods.
- Achieves comparable performance to fully supervised detectors with 12.8% less annotation.
- Surpasses fully supervised performance when using the entire annotation budget.

## Abstract

We study the problem of object detection from a novel perspective in which annotation budget constraints are taken into consideration, appropriately coined Budget Aware Object Detection (BAOD). When provided with a fixed budget, we propose a strategy for building a diverse and informative dataset that can be used to optimally train a robust detector. We investigate both optimization and learning-based methods to sample which images to annotate and what type of annotation (strongly or weakly supervised) to annotate them with. We adopt a hybrid supervised learning framework to train the object detector from both these types of annotation. We conduct a comprehensive empirical study showing that a handcrafted optimization method outperforms other selection techniques including random sampling, uncertainty sampling and active learning. By combining an optimal image/annotation selection scheme with hybrid supervised learning to solve the BAOD problem, we show that one can achieve the performance of a strongly supervised detector on PASCAL-VOC 2007 while saving 12.8% of its original annotation budget. Furthermore, when $100\%$ of the budget is used, it surpasses this performance by 2.0 mAP percentage points.