LSDA: Large Scale Detection Through Adaptation

Abstract

A major challenge in scaling object detection is the difficulty of obtaining labeled images for large numbers of categories. Recently, deep convolutional neural networks (CNNs) have emerged as clear winners on object classification benchmarks, in part due to training with 1.2M+ labeled classification images. Unfortunately, only a small fraction of those labels are available for the detection task. It is much cheaper and easier to collect large quantities of image-level labels from search engines than it is to collect detection data and label it with precise bounding boxes. In this paper, we propose Large Scale Detection through Adaptation (LSDA), an algorithm which learns the difference between the two tasks and transfers this knowledge to classifiers for categories without bounding box annotated data, turning them into detectors. Our method has the potential to enable detection for the tens of thousands of categories that lack bounding box annotations, yet have plenty of classification data. Evaluation on the ImageNet LSVRC-2013 detection challenge demonstrates the efficacy of our approach. This algorithm enables us to produce a >7.6K detector by using available classification data from leaf nodes in the ImageNet tree. We additionally demonstrate how to modify our architecture to produce a fast detector (running at 2fps for the 7.6K detector). Models and software are available at

References in corpus (4)

• Spatial Pyramid Pooling in Deep Convolutional Networks for Visual Recognition
• Descriptor Matching with Convolutional Neural Networks: a Comparison to SIFT
• Learning with Augmented Features for Heterogeneous Domain Adaptation
• On learning to localize objects with minimal supervision

Cited by in corpus (46)

• Learning Transferable Features with Deep Adaptation Networks
• Deep Visual Domain Adaptation: A Survey
• Unsupervised Domain Adaptation with Residual Transfer Networks
• Few-Example Object Detection with Model Communication
• Self Paced Deep Learning for Weakly Supervised Object Detection
• Recent Advances in Object Detection in the Age of Deep Convolutional Neural Networks
• Unsupervised Domain Adaptation with Adversarial Residual Transform Networks
• Visual and Semantic Knowledge Transfer for Large Scale Semi-supervised Object Detection
• Domain Adaptation for Object Detection via Style Consistency
• Multi-adversarial Faster-RCNN for Unrestricted Object Detection
• Learning to Transfer Examples for Partial Domain Adaptation
• SODA10M: A Large-Scale 2D Self/Semi-Supervised Object Detection Dataset for Autonomous Driving
• Unbiased Teacher for Semi-Supervised Object Detection
• Active and Incremental Learning with Weak Supervision
• Coupling Adversarial Learning with Selective Voting Strategy for Distribution Alignment in Partial Domain Adaptation
• Open-Vocabulary Object Detection Using Captions
• Towards Noise-resistant Object Detection with Noisy Annotations
• Associate-3Ddet: Perceptual-to-Conceptual Association for 3D Point Cloud Object Detection
• Proposal Learning for Semi-Supervised Object Detection
• 3DIoUMatch: Leveraging IoU Prediction for Semi-Supervised 3D Object Detection
• Learning from a Complementary-label Source Domain: Theory and Algorithms
• Gradient Regularized Contrastive Learning for Continual Domain Adaptation
• VISIR: Visual and Semantic Image Label Refinement
• Learning the semantic structure of objects from Web supervision
• Boosting Weakly Supervised Object Detection with Progressive Knowledge Transfer
• Mixed Supervised Object Detection with Robust Objectness Transfer
• Data-Uncertainty Guided Multi-Phase Learning for Semi-Supervised Object Detection
• Cross-Supervised Object Detection
• Learning to Localize Actions from Moments
• Scaling Object Detection by Transferring Classification Weights
• Adaptively-Accumulated Knowledge Transfer for Partial Domain Adaptation
• Domain Adaptive SiamRPN++ for Object Tracking in the Wild
• Exploring Bottom-up and Top-down Cues with Attentive Learning for Webly Supervised Object Detection
• Graphonomy: Universal Image Parsing via Graph Reasoning and Transfer
• MosaicOS: A Simple and Effective Use of Object-Centric Images for Long-Tailed Object Detection
• What leads to generalization of object proposals?
• Image Captioning with Unseen Objects
• Decoupled Gradient Harmonized Detector for Partial Annotation: Application to Signet Ring Cell Detection
• Range Adaptation for 3D Object Detection in LiDAR
• Adaptively Denoising Proposal Collection for Weakly Supervised Object Localization
• A Dictionary Approach to Domain-Invariant Learning in Deep Networks
• Domain Fingerprints for No-reference Image Quality Assessment
• Weak Novel Categories without Tears: A Survey on Weak-Shot Learning
• Unsupervised Object Detection with LiDAR Clues
• Unseen Object Segmentation in Videos via Transferable Representations
• CS-R-FCN: Cross-supervised Learning for Large-Scale Object Detection