Where are the objects in the image ?

Discover the current state of the art in objects detection.

Pascal VOC 2007 comp3 17 results collected

Units: mAP percent

Pascal VOC 2007 is commonly used because the test set has been realased. comp3 is the objects detection competition, using only the comp3 pascal training data.

Result	Method	Venue	Details
22.7 mAP	Ensemble of Exemplar-SVMs for Object Detection and Beyond	ICCV 2011	Details Reaches 19.8 mAP when not using context. Code available.
27.4 mAP	Measuring the objectness of image windows	PAMI 2012	Details Source code available.
28.7 mAP	Automatic discovery of meaningful object parts with latent CRFs	CVPR 2010
29.0 mAP	Object Detection with Discriminatively Trained Part Based Models	PAMI 2010	Details Original PAMI 2010 results, using voc-release3, with context. Obtains 26.2 mAP without context.
29.6 mAP	Latent Hierarchical Structural Learning for Object Detection	CVPR 2010
32.4 mAP	Deformable Part Models with Individual Part Scaling	BMVC 2013
34.3 mAP	Histograms of Sparse Codes for Object Detection	CVPR 2013	Details Reaches 26.9 mAP when using only the root templates (no parts).
34.3 mAP	Boosted local structured HOG-LBP for object localization	CVPR 2011
34.7 mAP	Discriminatively Trained And-Or Tree Models for Object Detection	CVPR 2013	Details The number of components, parts, their shape, appearance, and configuration are all learned.
34.7 mAP	Incorporating Structural Alternatives and Sharing into Hierarchy for Multiclass Object Recognition and Detection	CVPR 2013
34.8 mAP	Color Attributes for Object Detection	CVPR 2012	Details Code available.
35.4 mAP	Object Detection with Discriminatively Trained Part Based Models	PAMI 2010	Details 2012 results using voc-release5 (instead of release3 from 2010), with context. Obtains 33.7 mAP without context.
36.0 mAP	Machine Learning Methods for Visual Object Detection	archives-ouvertes 2011	Details Thesis results based on the features explained in our earlier paper “Feature Sets and Dimensionality Reduction For Visual Object Detection, BMVC, 2010”. Results table can be found on the page 104 of thesis.
38.7 mAP	Detection Evolution with Multi-Order Contextual Co-occurrence	CVPR 2013
40.5 mAP	Segmentation Driven Object Detection with Fisher Vectors	ICCV 2013	Details 5 mAP when not using inter-class contextual cues
41.7 mAP	Regionlets for Generic Object Detection	ICCV 2013	Details No context information being used.
43.7 mAP	Beyond Bounding-Boxes: Learning Object Shape by Model-Driven Grouping	ECCV 2012

Something is off, something is missing ? Feel free to fill in the form.

Pascal VOC 2007 comp4 3 results collected

Units: mAP percent

Just like comp3 but “any training data” can be used.

Result	Method	Venue	Details
59.2 mAP	Spatial Pyramid Pooling in Deep Convolutional Networks for Visual Recognition	ECCV 2014	Details 20x faster than R-CNN.
58.5 mAP	Rich feature hierarchies for accurate object detection and semantic segmentation	CVPR 2014
29.0 mAP	Multi-Component Models for Object Detection	ECCV 2012	Details 9 Monolithic, 29.0 Multi-components. Uses poselets and segmentation masks labels to build the clusters (thus comp4 not comp3).

Something is off, something is missing ? Feel free to fill in the form.

Pascal VOC 2010 comp3 10 results collected

Units: mAP percent

Pascal VOC 2010 version of the challenge. comp3 is the objects detection competition.

Result	Method	Venue	Details
24.98 mAP	Learning Collections of Part Models for Object Recognition	CVPR 2013
29.4 mAP	Discriminatively Trained And-Or Tree Models for Object Detection	CVPR 2013	Details The number of components, parts, their shape, appearance, and configuration are all learned.
33.4 mAP	Object Detection with Discriminatively Trained Part Based Models	PAMI 2010	Details This are the results of version 5 (September 2012) with context and extra octave. Raw detection obtains 29.6 mAP.
34.1 mAP	Segmentation as selective search for object recognition	ICCV 2011	Details Source code availabe at http://koen.me/research/selectivesearch/
35.1 mAP	Selective Search for Object Recognition	IJCV 2013
36.0 mAP	Latent Hierarchical Structural Learning for Object Detection	CVPR 2010	Details Result presented in the VOC 2010 workshop at ECCV 2010. They reported 35.99, which rounds up to 36.0.
36.8 mAP	Object Detection by Context and Boosted HOG-LBP	ECCV 2010	Details There seem to be no paper about this work, but only the slides from the entry at the PASCAL VOC workshop ECCV 2010.
38.4 mAP	Segmentation Driven Object Detection with Fisher Vectors	ICCV 2013	Details 8 mAP when not using inter-class contextual cues
39.7 mAP	Regionlets for Generic Object Detection	ICCV 2013	Details No context information being used.
40.4 mAP	Fisher and VLAD with FLAIR	CVPR 2014

Something is off, something is missing ? Feel free to fill in the form.

Pascal VOC 2010 comp4 3 results collected

Units: mAP percent

Just like comp3 but “any training data” can be used.

Result	Method	Venue	Details
53.7 mAP	Rich feature hierarchies for accurate object detection and semantic segmentation	CVPR 2014
40.4 mAP	Bottom-up Segmentation for Top-down Detection	CVPR 2013	Details Reaches 40.4 mAP when using full post-processing. Reaches 36.6 mAP without any post-processing. This method builds upon a sophisticated object segmentation method (CPMC).
33.1 mAP	Multi-Component Models for Object Detection	ECCV 2012	Details Multi-components result Uses poselets and segmentation masks labels to build the clusters (thus comp4 not comp3).

Something is off, something is missing ? Feel free to fill in the form.

Pascal VOC 2011 comp3

Units: mAP percent

Last Pascal VOC challenge instance (2012 version had identical data).

Result	Method	Venue	Details
40.6 mAP	Fisher and VLAD with FLAIR	CVPR 2014

Something is off, something is missing ? Feel free to fill in the form.