Detecting Moving Objects, Ghosts, and Shadows in Video Streams

By Rita Cucchiara, Costantino Grana, Massimo Piccardi, Andrea Prati. 2003

1,635 citations recorded by the source corpus.

Background subtraction methods are widely exploited for moving object detection in videos in many applications, such as traffic monitoring, human motion capture, and video surveillance. How to correctly and efficiently model and update the background model and how to deal with shadows are two of the most distinguishing and challenging aspects of such approaches. The article proposes a general-purpose method that combines statistical assumptions with the object-level knowledge of moving objects, apparent objects (ghosts), and shadows acquired in the processing of the previous frames. Pixels belonging to moving objects, ghosts, and shadows are processed differently in order to supply an object-based selective update. The proposed approach exploits color information for both background subtraction and shadow detection to improve object segmentation and background update. The approach proves fast, flexible, and precise in terms of both pixel accuracy and reactivity to background changes.

Subfields: Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

Citations in the corpus

On the Binding Problem in Artificial Neural Networks (2020, background): “…the form of a label) with each segment, and therefore can not segregate multiple objects belonging to the same class. that correspond to instances (provided that they move, e.g. Cucchiara et al., 2003), which can for example be learned through unsupervised multi-task learning (Ranjan et al., 2019).”
SAM-Adapter: Adapting Segment Anything in Underperformed Scenes (2023, background): “They can also negatively impact the performance of computer vision tasks [39, 9].”
DeshadowNet: A Multi-context Embedding Deep Network for Shadow Removal (2017, background): “The presence of illumination changes in an image, shadows in particular, have been proved to be one of the main challenging factors for a variety of computer vision tasks, such as object detection and tracking [6, 24].”
Comparative study of background subtraction algorithms (2010, methodology): “Other adaptation schemes have been proposed, some working at the pixel level, others at the blob level [10], and others being robust to shadows [11] Another method whose goal is to locally adapt to noise is the W 4 system by Haritaoglu et. al.”
A Multi-Task Mean Teacher for Semi-Supervised Shadow Detection (2020, neutral): “, image segmentation [4], object detection [2], and object tracking [27].”
Human Action Recognition With Video Data: Research and Evaluation Challenges (2014, neutral): “Artefacts such as shadows, complex backgrounds, ghosts and moving objects in the videos can affect shape based methods since they affect the performance of background subtraction [121].”
3DPeS: 3D people dataset for surveillance and forensics (2011, methodology): “); Tracking annotation will be generated using the SAKBOT system [5] as a baseline.”
A Self-Organizing Approach to Background Subtraction for Visual Surveillance Applications (2008, methodology): “Experimental results for moving object detection using the proposed approach have been produced for several image sequences.”
A Tensor-Based Information Framework for Predicting the Stock Market (2016, background): “Examples include detecting moving objects in video data [Cucchiara et al. 2003], context-aware mobile recommendations [Oku et al. 2006], and health care monitoring [Fleury et al. 2010].”
Shadow detection: A survey and comparative evaluation of recent methods (2012, methodology): “HSV [7], c1c2c3 [36], YUV [6], normalised RGB [4]), or a combination of them [41].”
Video Processing From Electro-Optical Sensors for Object Detection and Tracking in a Maritime Environment: A Survey (2016, background): “ployed for background suppression in [88], [89], [92], [119], [120].”
Fast Object Segmentation in Unconstrained Video (2013, methodology): “Our method does not require a static or slowly moving background (as opposed to classic background subtraction methods [9, 4, 7]).”
Moving Object Detection in Spatial Domain using Background Removal Techniques - State-of-Art (2008, methodology): “Another approach to introduce spatial variability is to use two thresholds with hysteresis [42, 93].”
A Review on Video-Based Human Activity Recognition (2013, background): “Another improvement is to discriminate moving objects, ghosts and shadow [40], based on statistical assumptions, with object-level knowledge, of moving objects, apparent objects (ghosts) and shadows.”
Visual Tracking: an Experimental Survey (2013, neutral): “For static scenes, background intensity representation is an old solution [84], [85] only suited for standardized circumstances, improved with background intensity prediction with simple statistics [86], [87], [88].”
Spatiotemporal Saliency in Dynamic Scenes (2010, methodology): “In computer vision, spatiotemporal saliency and the related task of background subtraction are commonly used as a preprocessing step for object and event detection [10], activity and gesture recognition [29], tracking [30], surveillance [12], and video retrieval [27].”
Performance evaluation of object detection algorithms for video surveillance (2006, methodology): “To deal with these difficulties several methods have been proposed (see [20]).”