Temporal Self-Ensembling Teacher for Semi-Supervised Object Detection (2020-07-13T00:00:00.000000Z)

TL;DR

A novel SSOD model called Temporal Self-Ensembling Teacher (TSET) is proposed, which ensembles its temporal predictions for unlabeled images under stochastic perturbations, and automatically reweights the inconsistent predictions, which preserves the knowledge for difficult objects to detect in the unlabeling images.

Abstract

This paper focuses on the semi-supervised object detection (SSOD) which makes good use of unlabeled data to boost performance. We face the following obstacles when adapting the knowledge distillation (KD) framework in SSOD. (1) The teacher model serves a dual role as a teacher and a student, such that the teacher predictions on unlabeled images may limit the upper bound of the student. (2) The data imbalance issue caused by the large quantity of consistent predictions between the teacher and student hinders an efficient knowledge transfer between them. To mitigate these issues, we propose a novel SSOD model called Temporal Self-Ensembling Teacher (TSET). Our teacher model ensembles its temporal predictions for unlabeled images under stochastic perturbations. Then, our teacher model ensembles its model weights with those of the student model by an exponential moving average. These ensembling strategies ensure data and model diversity, and lead to better teacher predictions for unlabeled images. In addition, we adapt the focal loss to formulate the consistency loss for handling the data imbalance issue. Together with a thresholding method, the focal loss automatically reweights the inconsistent predictions, which preserves the knowledge for difficult objects to detect in the unlabeled images. The mAP of our model reaches 80.73% and 40.52% on the VOC2007 test set and the COCO2014 minival5k set, respectively, and outperforms a strong fully supervised detector by 2.37% and 1.49%, respectively. Furthermore, the mAP of our model (80.73%) sets a new state-of-the-art performance in SSOD on the VOC2007 test set.

Authors

Cong Chen

1 papers

Shouyang Dong

1 papers

Ye Tian

1 papers

TL;DR

Abstract

Authors

References68 items

CKD: Cross-Task Knowledge Distillation for Text-to-Image Synthesis

A Simple Semi-Supervised Learning Framework for Object Detection

Reasoning on the Relation: Enhancing Visual Representation for Visual Question Answering and Cross-Modal Retrieval

Proposal Learning for Semi-Supervised Object Detection

International evaluation of an AI system for breast cancer screening

Momentum Contrast for Unsupervised Visual Representation Learning

Deep High-Resolution Representation Learning for Visual Recognition

Towards Understanding Knowledge Distillation

Scaling and Benchmarking Self-Supervised Visual Representation Learning

CenterNet: Keypoint Triplets for Object Detection

Structured Knowledge Distillation for Semantic Segmentation

Feature Selective Anchor-Free Module for Single-Shot Object Detection

Revisiting Self-Supervised Visual Representation Learning

Bottom-Up Object Detection by Grouping Extreme and Center Points

A Deep Learning Framework for Semi-Supervised Cross-Modal Retrieval with Label Prediction

Deep Learning for Generic Object Detection: A Survey

Leveraging Prior-Knowledge for Weakly Supervised Object Detection Under a Collaborative Self-Paced Curriculum Learning Framework

CornerNet: Detecting Objects as Paired Keypoints

Part-Aware Fine-Grained Object Categorization Using Weakly Supervised Part Detection Network

Min-Entropy Latent Model for Weakly Supervised Object Detection

The Power of Ensembles for Active Learning in Image Classification

Knowledge Distillation by On-the-Fly Native Ensemble

Deep Face Recognition: A Survey

YOLOv3: An Incremental Improvement

Data Distillation: Towards Omni-Supervised Learning

Cascade R-CNN: Delving Into High Quality Object Detection

DOTA: A Large-Scale Dataset for Object Detection in Aerial Images

Towards Effective Low-Bitwidth Convolutional Neural Networks

Squeeze-and-Excitation Networks

Multi-task Self-Supervised Visual Learning

Focal Loss for Dense Object Detection

Self-ensembling for visual domain adaptation

Deep Mutual Learning

Mask R-CNN

Weight-averaged consistency targets improve semi-supervised deep learning results

Weakly Supervised Learning of Deformable Part-Based Models for Object Detection via Region Proposals

YOLO9000: Better, Faster, Stronger

Feature Pyramid Networks for Object Detection

Multi-view 3D Object Detection Network for Autonomous Driving

Person Re-identification: Past, Present and Future

Temporal Ensembling for Semi-Supervised Learning

Regularization With Stochastic Transformations and Perturbations for Deep Semi-Supervised Learning

R-FCN: Object Detection via Region-based Fully Convolutional Networks

Inside-Outside Net: Detecting Objects in Context with Skip Pooling and Recurrent Neural Networks

Deep Residual Learning for Image Recognition

SSD: Single Shot MultiBox Detector

DenseBox: Unifying Landmark Localization with End to End Object Detection

Semi-supervised Learning with Ladder Networks

You Only Look Once: Unified, Real-Time Object Detection

Is object localization for free? - Weakly-supervised learning with convolutional neural networks

Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks

Fast R-CNN

Distilling the Knowledge in a Neural Network

Training Deep Neural Networks on Noisy Labels with Bootstrapping

From neural PCA to deep unsupervised learning

ImageNet Large Scale Visual Recognition Challenge

Microsoft COCO: Common Objects in Context

Effective Multiple Feature Hashing for Large-Scale Near-Duplicate Video Retrieval

Rich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation

Introduction to Semi-Supervised Learning

A Mutually Attentive Co-Training Framework for Semi-Supervised Recognition

Consistency-based Semi-supervised Learning for Object detection

Faster R-CNN and Retina network in PyTorch 1.0: Model zoo and baselines

Pseudo-Label : The Simple and Efficient Semi-Supervised Learning Method for Deep Neural Networks

Semi-Supervised Learning (Chapelle, O. et al., Eds.; 2006) [Book reviews]

Semi-Supervised Learning Literature Survey

The Open Images Dataset V4: Unified image classification, object detection, and visual relationship detection at scale

International Journal of Computer Vision manuscript No. (will be inserted by the editor) The PASCAL Visual Object Classes (VOC) Challenge

Field of Study

Journal Information

Name

Page

Volume

Venue Information

Name

Type