Detecting Moving Shadows: Algorithms and Evaluation

By Andrea Prati, Ivana Mikic, Mohan Manubhai Trivedi, Rita Cucchiara. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2003

864 citations recorded by the source corpus.

Moving shadows need careful consideration in the development of robust dynamic scene analysis systems. Moving shadow detection is critical for accurate object detection in video streams since shadow points are often misclassified as object points, causing errors in segmentation and tracking. Many algorithms have been proposed in the literature that deal with shadows. However, a comparative evaluation of the existing approaches is still lacking. In this paper, we present a comprehensive survey of moving shadow detection approaches. We organize contributions reported in the literature in four classes two of them are statistical and two are deterministic. We also present a comparative empirical evaluation of representative algorithms selected from these four classes. Novel quantitative (detection and discrimination rate) and qualitative metrics (scene and object independence, flexibility to shadow situations, and robustness to noise) are proposed to evaluate these classes of algorithms on a benchmark suite of indoor and outdoor video sequences. These video sequences and associated "ground-truth" data are made available at http://cvrr.ucsd.edu/aton/shadow to allow for others in the community to experiment with new algorithms and metrics.

Subfields: Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/6f0c3eaeca0872251a036b868f9002ad8f7cc045

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

Spatio-Temporal Data Mining (2017, background): “Existing image segmentation techniques that have been extended to work with videos to discover moving objects are relevant to address this problem [Haritaoglu et al. 2000; Prati et al. 2003].”
Background Subtraction in Real Applications: Challenges, Current Models and Future Directions (2019, background): “(2014) [114] SG [388] Idem SG Idem SG YCrCb Shadow detection [286] Quesada and Rodriguez (2016) [292] incPCP [305] Idem incPCP Idem incPCP Intensity -”
Cast shadow segmentation using invariant color features (2004, methodology): “In order to further evaluate the performance of the proposed algorithm, the video object extraction results obtained by combining the method described in [42] with the proposed shadow segmentation have been compared to four state-of-theart object extraction methods which are robust to shadows [45].”
Traditional and recent approaches in background modeling for foreground detection: An overview (2014, methodology): “From this dataset, five evaluated sequences can be downloaded: two of them (Laboratory, Highway) come from the ATON benchmark set [206], but the enclosed ground truth was generated by C.”
A comparative study on shadow compensation of color aerial images in invariant color models (2006, background): “The user’s accuracies are measures of the precision [25] of the algorithm and indicate the probabilities of pixels been correctly classified into actual categories on the ground.”
Pedestrian Protection Systems: Issues, Survey, and Challenges (2007, background): “A detailed comparative evaluation of various shadow detection algorithms in the literature is performed in [76].”
Shadow detection: A survey and comparative evaluation of recent methods (2012, methodology): “[31], many new methods have been proposed.”
Evaluation of background subtraction techniques for video surveillance (2011, methodology): “For the task of shadow removal, we apply the algorithm Cucchiara [7] proposed by Prati et al. [25] for general-purpose shadow detection with minimal assumptions.”
Efficient adaptive density estimation per image pixel for the task of background subtraction (2006, background): “Usually, we are interested only in the object and the pixels corresponding to the shadow should be detected (Prati et al., 2003).”
Physical models for moving shadow and object detection in video (2004, methodology): “Recently, statistical-based approaches using a mixture model have shown that many problematic phenomena such as repetitive motion of the background (e.g., swaying trees), sudden illumination changes (e.g., cloud cover), and sensor noise can be modeled [4], [8], [16].”
A texture-based method for modeling the background and detecting moving objects (2006, background): “Index Terms—Motion, texture, background subtraction, local binary pattern.”
Automatic Feature Learning for Robust Shadow Detection (2014, background): “Shadows are also a serious concern for aerial imaging and object tracking in video sequences [21].”
Segmentation and Tracking of Multiple Humans in Crowded Environments (2008, methodology): “A more sophisticated background model and shadow model (for example, [32]) could be used to improve the result.”
A comprehensive review of background subtraction algorithms evaluated with synthetic and real videos (2014, background): “Even if the evaluation of such algorithms is an important issue, only a few articles deal with this problematic in the literature [2,13,33].”
Modeling pixel process with scale invariant local patterns for background subtraction in complex scenes (2010, background): “A number of existing works handle illumination variation such as moving shadows in special color spaces [1] [7], but the learned parameters are not adaptable .”