Deep Domain Adaptation Based Video Smoke Detection using Synthetic Smoke Images

TL;DR

This paper introduces a deep domain adaptation method for video smoke detection that leverages synthetic smoke images to improve recognition accuracy on real images by reducing dataset bias.

Contribution

The paper presents a novel approach combining synthetic data generation and domain adaptation to enhance deep CNN performance in smoke detection tasks.

Findings

Synthetic smoke images improve training diversity.

Domain adaptation reduces feature distribution gap.

Enhanced model achieves higher recognition rates.

Abstract

In this paper, a deep domain adaptation based method for video smoke detection is proposed to extract a powerful feature representation of smoke. Due to the smoke image samples limited in scale and diversity for deep CNN training, we systematically produced adequate synthetic smoke images with a wide variation in the smoke shape, background and lighting conditions. Considering that the appearance gap (dataset bias) between synthetic and real smoke images degrades significantly the performance of the trained model on the test set composed fully of real images, we build deep architectures based on domain adaptation to confuse the distributions of features extracted from synthetic and real smoke images. This approach expands the domain-invariant feature space for smoke image samples. With their approximate feature distribution off non-smoke images, the recognition rate of the trained model is improved significantly compared to the model trained directly on mixed dataset of synthetic and real images. Experimentally, several deep architectures with different design choices are applied to the smoke detector. The ultimate framework can get a satisfactory result on the test set. We believe that our approach is a start in the direction of utilizing deep neural networks enhanced with synthetic smoke images for video smoke detection.