Improving object detection with deep convolutional networks via Bayesian optimization and structured prediction

TLDR

Deep CNN-based object detection has achieved breakthroughs, yet inaccurate localization remains a major source of error. The study seeks to improve localization accuracy in CNN-based object detection. The authors employ Bayesian optimization to propose bounding‑box candidates and train the CNN with a structured loss that penalizes localization errors. Experiments on PASCAL VOC 2007 and 2012 demonstrate that each proposed method improves detection over the baseline, and their combination significantly outperforms the previous state‑of‑the‑art.

Abstract

Object detection systems based on the deep convolutional neural network (CNN) have recently made ground- breaking advances on several object detection benchmarks. While the features learned by these high-capacity neural networks are discriminative for categorization, inaccurate localization is still a major source of error for detection. Building upon high-capacity CNN architectures, we address the localization problem by 1) using a search algorithm based on Bayesian optimization that sequentially proposes candidate regions for an object bounding box, and 2) training the CNN with a structured loss that explicitly penalizes the localization inaccuracy. In experiments, we demonstrated that each of the proposed methods improves the detection performance over the baseline method on PASCAL VOC 2007 and 2012 datasets. Furthermore, two methods are complementary and significantly outperform the previous state-of-the-art when combined.