Tensor Pooling Driven Instance Segmentation Framework for Baggage Threat Recognition

TL;DR

This paper introduces a novel multi-scale contour instance segmentation framework that effectively detects cluttered and concealed contraband in baggage X-ray images, outperforming existing methods on multiple public datasets.

Contribution

It proposes the first contour-based instance segmentation framework leveraging multi-scale information for baggage threat recognition in X-ray imagery.

Findings

Achieves state-of-the-art mean average precision scores on GDXray, SIXray, and OPIXray datasets.

Effectively handles high clutter, concealment, and occlusion in baggage X-ray scans.

First to utilize multi-scale contour information for this application.

Abstract

Automated systems designed for screening contraband items from the X-ray imagery are still facing difficulties with high clutter, concealment, and extreme occlusion. In this paper, we addressed this challenge using a novel multi-scale contour instance segmentation framework that effectively identifies the cluttered contraband data within the baggage X-ray scans. Unlike standard models that employ region-based or keypoint-based techniques to generate multiple boxes around objects, we propose to derive proposals according to the hierarchy of the regions defined by the contours. The proposed framework is rigorously validated on three public datasets, dubbed GDXray, SIXray, and OPIXray, where it outperforms the state-of-the-art methods by achieving the mean average precision score of 0.9779, 0.9614, and 0.8396, respectively. Furthermore, to the best of our knowledge, this is the first contour instance segmentation framework that leverages multi-scale information to recognize cluttered and concealed contraband data from the colored and grayscale security X-ray imagery.