Orientation Aware Object Detection with Application to Firearms




Automatic detection of firearms is important for enhancing the security and safety of people, however, it is a challenging task owing to the wide variations in shape, size, and appearance of firearms. Viewing angle variations and occlusions by the weapon’s carrier and the surrounding people, further increases the difficulty of the task. Moreover, the existing object detectors process rectangular areas, though a thin and long rifle may actually cover only a small percentage of that area and the rest may contain irrelevant details suppressing the required object signatures. To handle these challenges we propose an Orientation Aware Object Detector (OAOD) which has achieved improved firearm detection and localization performance. The proposed detector has two stages. In Stage-1 it predicts the orientation of the object which is used to rotate the object proposal. Maximum area rectangles are cropped from the rotated object proposals which are again classified and localized in Stage-2 of the algorithm. The oriented object proposals are mapped back to the original coordinates resulting in oriented bounding boxes that localize the weapons much better than the axis-aligned bounding boxes. Being orientation aware, our non-maximum suppression is able to avoid multiple detections of the same object and it can better resolve objects which lie in close proximity to each other. This two-stage system leverages OAOD to predict object-oriented bounding boxes while being trained only on the axis-aligned boxes in the ground-truth. In order to train object detectors for firearm detection, a dataset consisting of around 11,000 firearm images are collected from the internet and manually annotated. The proposed ITU Firearms (ITUF) dataset contains a wide range of guns and rifles. The OAOD algorithm is evaluated on the ITUF dataset and compared with the current state of the art object detectors. Our experiments demonstrate the excellent performance of the proposed detector for the task of firearm detection.


Key Idea:

One limitation of the existing methods is the use of axis-aligned windows for object detection. Most classifiers decide the presence of an object by analyzing the features in that window. The physically thin and elongated structure of most of the rifles and small size of most guns, make these axis-aligned windows inefficient due to low signal to noise ratio where the signal is the firearm signature and noise is everything else in the window. In the case of firearms being carried by a person, the window will tend to contain substantial information belonging to the background or non-firearm objects, like the person himself. This mixture of the information makes it difficult for classifiers to learn to separate the required information or signal from the other objects acting as noise.


Key idea proposed to handle this issue is illustrated below. By having the the maximum area rectangle and making oriented region proposal (ORP) containing only firearms.


Proposed OAOD Overview:

Most of the current object detectors employ axis-aligned bounding boxes which may incur noise and clutter due to uncorrelated background objects. Such noise may adversely effect the performance of object detection. To overcome this issue, we propose Orientation Aware Object Detection (OAOD) algorithm which consists of a single pipe-lined network consisting of a cascade of two stages shown below. In the Stage-1, the orientation of the object is predicted along with the classification score and offset to the region proposal. In the Stage-2, the updated object proposals are warped according to the predicted orientation such that the object becomes axis aligned. The maximum area rectangle contained within the rotated proposal is cropped and the final Oriented Region Proposal is used for further classification and offset regression. Thus the redundant information contained within the region proposals is significantly reduced for the case of non-axis aligned firearms. The classifier in Stage-2 is trained to predict classification scores on these oriented region proposals to achieve better performance. The proposed network takes an entire image as input and detects and localizes two types of firearms including rifles and guns.



Dataset (DATASET is available upon request [Google Form])

Orientation wise dataset distribution and sample images from dataset.



arXiv ( Paper: OAOD with Application to Firearms )

Github (Github: OAOD with Application to Firearms)



Object Detection, Orientation, Firearms Detection, Orientation Aware Firearms Detection, Orientation Aware Object Detection, Oriented Object Detection, OAOD, Deep Learning, Computer Vision, Orientation Aware Object Detection with Application to Firearms



author = {Javed Iqbal and Muhammad Akhtar Munir and Arif Mahmood and Afsheen Rafaqat Ali and Mohsen Ali},
title = {Orientation Aware Object Detection with Application to Firearms},
journal = {CoRR},
volume = {abs/1904.10032},
year = {2019},
url = {http://arxiv.org/abs/1904.10032} }