Focal Loss for Dense Object Detection
Reads0
Chats0
TLDR
Focal loss as discussed by the authors focuses training on a sparse set of hard examples and prevents the vast number of easy negatives from overwhelming the detector during training, which improves the accuracy of one-stage detectors.Abstract:
The highest accuracy object detectors to date are based on a two-stage approach popularized by R-CNN, where a classifier is applied to a sparse set of candidate object locations. In contrast, one-stage detectors that are applied over a regular, dense sampling of possible object locations have the potential to be faster and simpler, but have trailed the accuracy of two-stage detectors thus far. In this paper, we investigate why this is the case. We discover that the extreme foreground-background class imbalance encountered during training of dense detectors is the central cause. We propose to address this class imbalance by reshaping the standard cross entropy loss such that it down-weights the loss assigned to well-classified examples. Our novel Focal Loss focuses training on a sparse set of hard examples and prevents the vast number of easy negatives from overwhelming the detector during training. To evaluate the effectiveness of our loss, we design and train a simple dense detector we call RetinaNet. Our results show that when trained with the focal loss, RetinaNet is able to match the speed of previous one-stage detectors while surpassing the accuracy of all existing state-of-the-art two-stage detectors. Code is at: https://github.com/facebookresearch/Detectron .read more
Citations
More filters
Posted Content
Preparing Lessons: Improve Knowledge Distillation with Better Supervision
TL;DR: In this paper, the authors introduce two novel approaches, Knowledge Adjustment (KA) and Dynamic Temperature Distillation (DTD), to penalize bad supervision and improve student model, which can get encouraging performance compared with state-of-the-art methods.
Posted Content
Graph Density-Aware Losses for Novel Compositions in Scene Graph Generation
Boris Knyazev,Harm de Vries,Catalina Cangea,Graham W. Taylor,Aaron Courville,Eugene Belilovsky +5 more
TL;DR: A density-normalized edge loss is introduced, which provides more than a two-fold improvement in certain generalization metrics in scene graph generation, and highlights the difficulty of accurately evaluating models using existing metrics, especially on zero/few shots, and introduces a novel weighted metric.
Book ChapterDOI
Ischemic Stroke Lesion Segmentation in CT Perfusion Scans Using Pyramid Pooling and Focal Loss
TL;DR: In this article, a fully convolutional neural network was used for segmenting ischemic stroke lesions in CT perfusion images for the ISLES 2018 challenge, which was based on the PSPNet.
Journal ArticleDOI
Vision-Based Fall Event Detection in Complex Background Using Attention Guided Bi-Directional LSTM
TL;DR: Evaluation of the algorithm performances in comparison with other state-of-the-art methods indicates that the proposed design is accurate and robust, which means it is suitable for the task of fall event detection in complex situation.
Journal ArticleDOI
Using Vehicle Synthesis Generative Adversarial Networks to Improve Vehicle Detection in Remote Sensing Images
TL;DR: A learning method named Vehicle Synthesis Generative Adversarial Networks (VS-GANs) to generate annotated vehicles from remote sensing images that significantly improves the generalization capability of vehicle detectors.
References
More filters
Proceedings ArticleDOI
Deep Residual Learning for Image Recognition
TL;DR: In this article, the authors proposed a residual learning framework to ease the training of networks that are substantially deeper than those used previously, which won the 1st place on the ILSVRC 2015 classification task.
Proceedings Article
ImageNet Classification with Deep Convolutional Neural Networks
TL;DR: The state-of-the-art performance of CNNs was achieved by Deep Convolutional Neural Networks (DCNNs) as discussed by the authors, which consists of five convolutional layers, some of which are followed by max-pooling layers, and three fully-connected layers with a final 1000-way softmax.
Proceedings ArticleDOI
Histograms of oriented gradients for human detection
Navneet Dalal,Bill Triggs +1 more
TL;DR: It is shown experimentally that grids of histograms of oriented gradient (HOG) descriptors significantly outperform existing feature sets for human detection, and the influence of each stage of the computation on performance is studied.
Book ChapterDOI
Microsoft COCO: Common Objects in Context
Tsung-Yi Lin,Michael Maire,Serge Belongie,James Hays,Pietro Perona,Deva Ramanan,Piotr Dollár,C. Lawrence Zitnick +7 more
TL;DR: A new dataset with the goal of advancing the state-of-the-art in object recognition by placing the question of object recognition in the context of the broader question of scene understanding by gathering images of complex everyday scenes containing common objects in their natural context.
Proceedings ArticleDOI
Fully convolutional networks for semantic segmentation
TL;DR: The key insight is to build “fully convolutional” networks that take input of arbitrary size and produce correspondingly-sized output with efficient inference and learning.