FADE: A Dataset for Detecting Falling Objects around Buildings in Video

2024-08-11

arXiv.org

AI Summary

Falling objects near buildings pose significant challenges for real-time detection due to their high impact velocity, small size, rapid motion, and cluttered backgrounds—rendering manual surveillance ineffective. To address this, we introduce FADE, the first large-scale video dataset specifically designed for this scenario, comprising 1,881 multi-scene, multi-weather, and multi-resolution videos. We further propose FADE-Net, a dedicated detection architecture that leverages optical-flow-guided motion modeling to enhance responses to small, fast-moving objects, integrates multi-scale spatiotemporal features, and adopts a lightweight design for efficiency. This work formally defines “falling-object detection near buildings” as a novel task paradigm. Extensive experiments on FADE demonstrate that FADE-Net significantly outperforms state-of-the-art general object detectors, video object detectors, and motion-based detectors, establishing a new benchmark. Both the FADE dataset and FADE-Net code are publicly released.

Technology Category

• Computer Vision: Object Detection & Categorization
• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Natural Language Processing: Fact-Checking / Misinformation Detection (NLP Focus)

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Energy management for devices in mobile Web and WoT environments
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web

Abstract

Falling objects from buildings can cause severe injuries to pedestrians due to the great impact force they exert. Although surveillance cameras are installed around some buildings, it is challenging for humans to capture such events in surveillance videos due to the small size and fast motion of falling objects, as well as the complex background. Therefore, it is necessary to develop methods to automatically detect falling objects around buildings in surveillance videos. To facilitate the investigation of falling object detection, we propose a large, diverse video dataset called FADE (FAlling Object DEtection around Buildings) for the first time. FADE contains 1,881 videos from 18 scenes, featuring 8 falling object categories, 4 weather conditions, and 4 video resolutions. Additionally, we develop a new object detection method called FADE-Net, which effectively leverages motion information and produces small-sized but high-quality proposals for detecting falling objects around buildings. Importantly, our method is extensively evaluated and analyzed by comparing it with the previous approaches used for generic object detection, video object detection, and moving object detection on the FADE dataset. Experimental results show that the proposed FADE-Net significantly outperforms other methods, providing an effective baseline for future research. The dataset and code are publicly available at https://fadedataset.github.io/FADE.github.io/.

Problem

• Detecting small, fast-moving falling objects in surveillance videos
• Addressing complex backgrounds in building surveillance footage
• Automating falling object detection to prevent pedestrian injuries

Innovation

• Proposes FADE-Net for motion-based detection
• Generates small high-quality object proposals
• Leverages motion information effectively