Communications of the ACM
Gunfire or Plastic Bag Popping? Trained Computer Knows the Difference
The results of the study demonstrated how fake gunshot sounds can confuse a gunshot sound detection system.
Credit: Florida Atlantic University News Desk
Florida Atlantic University (FAU) researchers studied the reliability of acoustic gunshot sound detector systems based on a dataset of audio recordings of plastic bag explosions in different environments and conditions.
They found that gunshot sound detector systems misclassified 75% of plastic bag popping sounds as gunshot sounds.
The researchers also developed a classification algorithm based on a convolutional neural network and trained a model to differentiate between gunshots and plastic bag explosions.
The model was trained on 374 gunshot samples from an urban sound database.
The researchers found the model performed well in distinguishing gunshots from plastic bags popping.
Said FAU's Hanqi Zhuang, "We used different environments to give the machine learning algorithm a better perception sense of the differentiation of the closely related sounds."
From Florida Atlantic University News Desk
View Full Article
Abstracts Copyright © 2021 SmithBucklin, Washington, DC, USA
No entries found