Face Detection using Deep Learning: An Improved Faster RCNN Approach

By Xudong Sun, Pengcheng Wu, Steven C. H. Hoi. Neurocomputing, 2017

661 citations recorded by the source corpus.

Abstract In this paper, we present a new face detection scheme using deep learning and achieve the state-of-the-art detection performance on the well-known FDDB face detection benchmark evaluation. In particular, we improve the state-of-the-art Faster RCNN framework by combining a number of strategies, including feature concatenation, hard negative mining, multi-scale training, model pre-training, and proper calibration of key parameters. As a consequence, the proposed scheme obtained the state-of-the-art face detection performance and was ranked as one of the best models in terms of ROC curves of the published methods on the FDDB benchmark. 1

Subfields: Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/84dcf04802743d9907b5b3ae28b19cbbacd97981

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

S^3FD: Single Shot Scale-Invariant Face Detector (2017, methodology): “Finally, we evaluate our face detector on FDDB against the state-of-the-art methods [1, 6, 7, 15, 18, 19, 20, 22, 24, 25, 31, 35, 36, 43, 47, 49, 55, 57, 58].”
When Machine Learning Meets Privacy (2020, background): “ck tool here. •The re-identification attack can be launched by face recognition techniques. The recent advanceinDNNmakesitmoreharmfulfromtwoaspects.First,theprocessbecomesautomatic with high accuracy [67, 151, 168]. Second, traditional protection schemes such as obfuscation no longer work effectivel”
Recent Advances in Deep Learning for Object Detection (2019, methodology): “[183] proposed a Fast R-CNN based framework which integrated multiscale features for prediction and converted the resulting detection bounding boxes into ellipses as the regions of human faces are more elliptical than rectangular.”
Object Detection With Deep Learning: A Review (2018, background): “From Figure 13(b), it can be observed that most of CNN based methods earn similar true positive rates between 60% and 70% while DeepIR and HR-ER perform much better than them.”
Unsupervised Hard Example Mining from Videos for Improved Object Detection (2018, methodology): “We crucially differ here from recent bootstrapping approaches [55,51] by (a) using large amounts of unlabeled data available on the web instead of relying only on the limited fully-supervised training data from WIDER Face [61] or Caltech Pedestrians [12], and (b) having a novel filtering criterion o”
Towards a Visual Privacy Advisor: Understanding and Predicting Privacy Risks in Images (2017, neutral): “Privacy and Computer Vision Several works explore detecting individual privacy attributes such as license plates [53, 49, 6], age estimation from facial photographs [4], social relationships [45], face detection [40, 44], landmark detection [51] and occupation recognition [39].”