Results 1 -
3 of
3
Deepid-net: Deformable deep convolutional neural networks for object detection
- In CVPR
, 2015
"... In this paper, we propose deformable deep convolutional neural networks for generic object detection. This new deep learning object detection framework has innovations in multiple aspects. In the proposed new deep architecture, a new deformation constrained pooling (def-pooling) layer models the def ..."
Abstract
-
Cited by 3 (3 self)
- Add to MetaCart
(Show Context)
In this paper, we propose deformable deep convolutional neural networks for generic object detection. This new deep learning object detection framework has innovations in multiple aspects. In the proposed new deep architecture, a new deformation constrained pooling (def-pooling) layer models the deformation of object parts with geometric con-straint and penalty. A new pre-training strategy is proposed to learn feature representations more suitable for the object detection task and with good generalization capability. By changing the net structures, training strategies, adding and removing some key components in the detection pipeline, a set of models with large diversity are obtained, which significantly improves the effectiveness of model averag-ing. The proposed approach improves the mean averaged precision obtained by RCNN [14], which was the state-of-the-art, from 31 % to 50.3 % on the ILSVRC2014 detection test set. It also outperforms the winner of ILSVRC2014, GoogLeNet, by 6.1%. Detailed component-wise analysis is also provided through extensive experimental evaluation, which provide a global view for people to understand the deep learning object detection pipeline. 1.
Minimum Barrier Salient Object Detection at 80 FPS
"... Sample saliency maps of several state-of-the-art methods (SO [39], AMC [15], HS [34] and SIA [6]) and methods with fast speed (HC [5], FT [1] and ours). Our method runs at about 80 FPS using a single thread, and produces saliency maps of high quality. Previous methods with similar speed, like HC and ..."
Abstract
- Add to MetaCart
(Show Context)
Sample saliency maps of several state-of-the-art methods (SO [39], AMC [15], HS [34] and SIA [6]) and methods with fast speed (HC [5], FT [1] and ours). Our method runs at about 80 FPS using a single thread, and produces saliency maps of high quality. Previous methods with similar speed, like HC and FT, usually cannot handle complex images well. We propose a highly efficient, yet powerful, salient object detection method based on the Minimum Barrier Distance (MBD) Transform. The MBD transform is robust to pixel-value fluctuation, and thus can be effectively applied on raw pixels without region abstraction. We present an approx-imate MBD transform algorithm with 100X speedup over the exact algorithm. An error bound analysis is also pro-vided. Powered by this fast MBD transform algorithm, the proposed salient object detection method runs at 80 FPS, and significantly outperforms previous methods with similar speed on four large benchmark datasets, and achieves com-parable or better performance than state-of-the-art meth-ods. Furthermore, a technique based on color whitening is proposed to extend our method to leverage the appearance-based backgroundness cue. This extended version further improves the performance, while still being one order of magnitude faster than all the other leading methods. 1.
Partial Occlusion Handling in Pedestrian Detection with a Deep Model
, 2012
"... Part-based models have demonstrated their merit in object detection. However, there is a key issue to be solved on how to integrate the inaccurate scores of part detectors when there are occlusions, abnormal deformations, appearances or illuminations. To handle the imperfection of part detectors, t ..."
Abstract
- Add to MetaCart
(Show Context)
Part-based models have demonstrated their merit in object detection. However, there is a key issue to be solved on how to integrate the inaccurate scores of part detectors when there are occlusions, abnormal deformations, appearances or illuminations. To handle the imperfection of part detectors, this paper presents a probabilistic pedestrian detection framework. In this framework, a deformable part-based model is used to obtain the scores of part detectors and the visibilities of parts are modeled as hidden variables. Once the occluded parts are identified, their effects are properly removed from the final detection score. Unlike previous occlusion handling approaches that assumed independence among the visibility probabilities of parts or manually defined rules for the visibility relationship, a deep model is proposed in this paper for learning the visibility relationship among overlapping parts at multiple layers. The proposed approach can be viewed as a general post-processing of part-detection results and can take detection scores of existing part-based models as input. Experimental results on three public datasets (Caltech, ETH and Daimler) and a new CUHK occlusion dataset1, which is specially designed for the evaluation of occlusion handling approaches, show the effectiveness of the proposed approach.
