Efficient Sample Mining for Object Detection

Olivier Canevet, Francois Fleuret
Proceedings of the Sixth Asian Conference on Machine Learning, PMLR 39:48-63, 2015.

Abstract

Object detectors based on the sliding window technique are usually trained in two successive steps: first, an initial classifier is trained on a population of positive samples (i.e. images of the object to detect) and negative samples randomly extracted from scenes which do not contain the object to detect. Then, the scenes are scanned with that initial classifier to enrich the initial set with negative samples incorrectly classified as positive. This bootstrapping process provides the learning algorithm with "hard" samples, which help to improve the decision boundary. Little work has been done on how to efficiently enrich the training set. While the standard bootstrapping approach densely visits the scenes, we propose to evaluate which regions of scenes can be discarded without any further computation to concentrate the search on promising areas. We apply our method to two standard object detection settings, pedestrian and face detection, and show that it provides a multi-fold speed up.

Cite this Paper


BibTeX
@InProceedings{pmlr-v39-canevet14a, title = {Efficient Sample Mining for Object Detection}, author = {Canevet, Olivier and Fleuret, Francois}, booktitle = {Proceedings of the Sixth Asian Conference on Machine Learning}, pages = {48--63}, year = {2015}, editor = {Phung, Dinh and Li, Hang}, volume = {39}, series = {Proceedings of Machine Learning Research}, address = {Nha Trang City, Vietnam}, month = {26--28 Nov}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v39/canevet14a.pdf}, url = {https://proceedings.mlr.press/v39/canevet14a.html}, abstract = {Object detectors based on the sliding window technique are usually trained in two successive steps: first, an initial classifier is trained on a population of positive samples (i.e. images of the object to detect) and negative samples randomly extracted from scenes which do not contain the object to detect. Then, the scenes are scanned with that initial classifier to enrich the initial set with negative samples incorrectly classified as positive. This bootstrapping process provides the learning algorithm with "hard" samples, which help to improve the decision boundary. Little work has been done on how to efficiently enrich the training set. While the standard bootstrapping approach densely visits the scenes, we propose to evaluate which regions of scenes can be discarded without any further computation to concentrate the search on promising areas. We apply our method to two standard object detection settings, pedestrian and face detection, and show that it provides a multi-fold speed up.} }
Endnote
%0 Conference Paper %T Efficient Sample Mining for Object Detection %A Olivier Canevet %A Francois Fleuret %B Proceedings of the Sixth Asian Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2015 %E Dinh Phung %E Hang Li %F pmlr-v39-canevet14a %I PMLR %P 48--63 %U https://proceedings.mlr.press/v39/canevet14a.html %V 39 %X Object detectors based on the sliding window technique are usually trained in two successive steps: first, an initial classifier is trained on a population of positive samples (i.e. images of the object to detect) and negative samples randomly extracted from scenes which do not contain the object to detect. Then, the scenes are scanned with that initial classifier to enrich the initial set with negative samples incorrectly classified as positive. This bootstrapping process provides the learning algorithm with "hard" samples, which help to improve the decision boundary. Little work has been done on how to efficiently enrich the training set. While the standard bootstrapping approach densely visits the scenes, we propose to evaluate which regions of scenes can be discarded without any further computation to concentrate the search on promising areas. We apply our method to two standard object detection settings, pedestrian and face detection, and show that it provides a multi-fold speed up.
RIS
TY - CPAPER TI - Efficient Sample Mining for Object Detection AU - Olivier Canevet AU - Francois Fleuret BT - Proceedings of the Sixth Asian Conference on Machine Learning DA - 2015/02/16 ED - Dinh Phung ED - Hang Li ID - pmlr-v39-canevet14a PB - PMLR DP - Proceedings of Machine Learning Research VL - 39 SP - 48 EP - 63 L1 - http://proceedings.mlr.press/v39/canevet14a.pdf UR - https://proceedings.mlr.press/v39/canevet14a.html AB - Object detectors based on the sliding window technique are usually trained in two successive steps: first, an initial classifier is trained on a population of positive samples (i.e. images of the object to detect) and negative samples randomly extracted from scenes which do not contain the object to detect. Then, the scenes are scanned with that initial classifier to enrich the initial set with negative samples incorrectly classified as positive. This bootstrapping process provides the learning algorithm with "hard" samples, which help to improve the decision boundary. Little work has been done on how to efficiently enrich the training set. While the standard bootstrapping approach densely visits the scenes, we propose to evaluate which regions of scenes can be discarded without any further computation to concentrate the search on promising areas. We apply our method to two standard object detection settings, pedestrian and face detection, and show that it provides a multi-fold speed up. ER -
APA
Canevet, O. & Fleuret, F.. (2015). Efficient Sample Mining for Object Detection. Proceedings of the Sixth Asian Conference on Machine Learning, in Proceedings of Machine Learning Research 39:48-63 Available from https://proceedings.mlr.press/v39/canevet14a.html.

Related Material