Improving an Object Detector and Extracting Regions Using Superpixels

Shu, Guang; Dehghan, Afshin; Shah, Mubarak

doi:10.1109/cvpr.2013.477

Cited by 67 publications

(52 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…As can be observed in Section 4.3, the overwhelming majority of the state-of-the-art research for domain adaptation of object detectors in videos use self-training in one form or another [29,30,31,32,33,36,37,38,40,42,11,10]. In order to adapt a generic pedestrian detector to a specific scene, a typical system would run the generic detector on some frames in a video, then score each detection using some heuristics and afterwards, add the most confident positive and negative detections to the original dataset for retraining.…”

Section: Discussionmentioning

confidence: 99%

“…at least two) source domains whereas most of the works reviewed in this paper assume that only one source domain is available. The approach of Shu et al [42] may work poorly for videos with small pedestrians.…”

Section: Discussionmentioning

confidence: 99%

“…Shu et al [42] propose a self-training approach to adapt a generic pedestrian to specific videos. Firstly, the generic detector is applied on frames and then the most confident detections are collected as positive examples.…”

Section: Domain Adaptation For Object Detection In Videosmentioning

confidence: 99%

See 2 more Smart Citations

Adapting pedestrian detectors to new domains: A comprehensive review

Htike

Hogg

2016

Engineering Applications of Artificial Intelligence

View full text Add to dashboard Cite

Successful detection and localisation of pedestrians is an important goal in computer vision which is a core area in Artificial Intelligence. State-of-the-art pedestrian detectors proposed in literature have reached impressive performance on certain datasets. However, it has been pointed out that these detectors tend not to perform very well when applied to specific scenes that differ from the training datasets in some ways. Due to this, domain adaptation approaches have recently become popular in order to adapt existing detectors to new domains to improve the performance in those domains. There is a real need to review and analyse critically the state-of-the-art domain adaptation algorithms, especially in the area of object and pedestrian detection. In this paper, we survey the most relevant and important state-of-the-art results for domain adaptation for image and video data, with a particular focus on pedestrian detection. Related areas to domain adaptation are also included in our review and we make observations and draw conclusions from the representative papers and give practical recommendations on which methods should be preferred in different situations that practitioners may encounter in real-life.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Adapting pedestrian detectors to new domains: A comprehensive review

Htike

Hogg

2016

Engineering Applications of Artificial Intelligence

View full text Add to dashboard Cite

show abstract

“…Instead, our goal is to train the best object detectors for a specific scenario. More recently, [20,21,22] improve generic offline trained detectors using specific scenarios.…”

Section: Related Workmentioning

confidence: 99%

Combining where and what in change detection for unsupervised foreground learning in surveillance

Huerta

Pedersoli

Gonzàlez³

et al. 2015

Pattern Recognition

View full text Add to dashboard Cite

Change detection is the most important task for video surveillance analytics such as foreground and anomaly detection. Current foreground detectors learn models from annotated images since the goal is to generate a robust foreground model able to detect changes in all possible scenarios. Unfortunately, manual labelling is very expensive. Most advanced supervised learning techniques based on generic object detection datasets currently exhibit very poor performance when applied to surveillance datasets because of the unconstrained nature of such environments in terms of types and appearances of objects. In this paper, we take advantage of change detection for training multiple foreground detectors in an unsupervised manner. We use statistical learning techniques which exploit the use of latent parameters for selecting the best foreground model parameters for a given scenario. In essence, the main novelty of our proposed approach is to combine the where (motion segmentation) and what (learning procedure) in change detection in Preprint submitted to Pattern Recognition November 3, 2014 an unsupervised way for improving the specificity and generalization power of foreground detectors at the same time. We propose a framework based on latent Support Vector Machines that, given a noisy initialization based on motion cues, learns the correct position, aspect ratio, and appearance of all moving objects in a particular scene. Specificity is achieved by learning the particular change detections of a given scenario, and generalization is guaranteed since our method can be applied to any possible scene and foreground object, as demonstrated in the experimental results outperforming the state-of-the-art.

show abstract

“…In a comprehensive study, Dollár et al (2011) show that the recall rates of 16 different pedestrian detectors decrease rapidly if the intersection-over-union score threshold is increased. A better alignment of the detection result to the real object boundaries is for instance achieved by finer segmentation, based on pixels (Dai and Hoiem, 2012), superpixels (Shu et al, 2013), interest points (Ommer et al, 2009), (Gall and Lempitsky, 2013), object parts (Felzenszwalb et al, 2010), (Benfold and Reid, 2011) or contour models (Leibe et al, 2005), (Gavrila and Munder, 2007). Such models have the advantage of being more robust against partial occlusions compared to a holistic model.…”

Section: Introductionmentioning

confidence: 99%

Probabilistic Multi-Person Tracking Using Dynamic Bayes Networks

Klinger

Rottensteiner

Heipke

2015

ISPRS Ann. Photogramm. Remote Sens. Spatial Inf. Sci.

View full text Add to dashboard Cite

ABSTRACT:Tracking-by-detection is a widely used practice in recent tracking systems. These usually rely on independent single frame detections that are handled as observations in a recursive estimation framework. If these observations are imprecise the generated trajectory is prone to be updated towards a wrong position. In contrary to existing methods our novel approach uses a Dynamic Bayes Network in which the state vector of a recursive Bayes filter, as well as the location of the tracked object in the image are modelled as unknowns. These unknowns are estimated in a probabilistic framework taking into account a dynamic model, and a state-of-the-art pedestrian detector and classifier. The classifier is based on the Random Forest-algorithm and is capable of being trained incrementally so that new training samples can be incorporated at runtime. This allows the classifier to adapt to the changing appearance of a target and to unlearn outdated features. The approach is evaluated on a publicly available benchmark. The results confirm that our approach is well suited for tracking pedestrians over long distances while at the same time achieving comparatively good geometric accuracy.

show abstract

Improving an Object Detector and Extracting Regions Using Superpixels

Cited by 67 publications

References 14 publications

Adapting pedestrian detectors to new domains: A comprehensive review

Adapting pedestrian detectors to new domains: A comprehensive review

Combining where and what in change detection for unsupervised foreground learning in surveillance

Probabilistic Multi-Person Tracking Using Dynamic Bayes Networks

Contact Info

Product

Resources

About