Weakly Supervised Object Localization with Multi-fold Multiple Instance Learning
arXiv:1503.00949 · doi:10.1109/TPAMI.2016.2535231
Abstract
Object category localization is a challenging problem in computer vision. Standard supervised training requires bounding box annotations of object instances. This time-consuming annotation process is sidestepped in weakly supervised learning. In this case, the supervised information is restricted to binary labels that indicate the absence/presence of object instances in the image, without their locations. We follow a multiple-instance learning approach that iteratively trains the detector and infers the object locations in the positive training images. Our main contribution is a multi-fold multiple instance learning procedure, which prevents training from prematurely locking onto erroneous object locations. This procedure is particularly important when using high-dimensional representations, such as Fisher vectors and convolutional neural network features. We also propose a window refinement method, which improves the localization accuracy by incorporating an objectness prior. We present a detailed experimental evaluation using the PASCAL VOC 2007 dataset, which verifies the effectiveness of our approach.
To appear in IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI)
References in corpus (2)
Cited by in corpus (107)
- Grad-CAM: Visual Explanations from Deep Networks via Gradient-based Localization
- Multiple Instance Learning: A Survey of Problem Characteristics and Applications
- Object Detection in 20 Years: A Survey
- Deep Learning for Generic Object Detection: A Survey
- Learning Deep Features for Discriminative Localization
- Constrained Deep Weak Supervision for Histopathology Image Segmentation
- HA-CCN: Hierarchical Attention-based Crowd Counting Network
- Learning to detect chest radiographs containing lung nodules using visual attention networks
- Self Paced Deep Learning for Weakly Supervised Object Detection
- Visual Search at Alibaba
- Human Action Localization with Sparse Spatial Supervision
- Hide-and-Seek: A Data Augmentation Technique for Weakly-Supervised Localization and Beyond
- Domain Adaptation for Object Detection via Style Consistency
- Weakly Supervised Object Detection in Artworks
- Toward unsupervised, multi-object discovery in large-scale image collections
- Deep Weakly-Supervised Learning Methods for Classification and Localization in Histology Images: A Survey
- ContextLocNet: Context-Aware Deep Network Models for Weakly Supervised Localization
- Dual-stream Multiple Instance Learning Network for Whole Slide Image Classification with Self-supervised Contrastive Learning
- Two-Phase Learning for Weakly Supervised Object Localization
- Learning Inductive Attention Guidance for Partially Supervised Pancreatic Ductal Adenocarcinoma Prediction
- UntrimmedNets for Weakly Supervised Action Recognition and Detection
- C-MIL: Continuation Multiple Instance Learning for Weakly Supervised Object Detection
- Multiple Instance Detection Network with Online Instance Classifier Refinement
- Deep Self-Taught Learning for Weakly Supervised Object Localization
- Hierarchical Context Embedding for Region-based Object Detection
- Soft Proposal Networks for Weakly Supervised Object Localization
- Sharing Pain: Using Pain Domain Transfer for Video Recognition of Low Grade Orthopedic Pain in Horses
- Weakly and Semi Supervised Detection in Medical Imaging via Deep Dual Branch Net
- Weakly-supervised Visual Grounding of Phrases with Linguistic Structures
- Backtracking Spatial Pyramid Pooling (SPP)-based Image Classifier for Weakly Supervised Top-down Salient Object Detection
- Pointly-Supervised Action Localization
- PPR-FCN: Weakly Supervised Visual Relation Detection via Parallel Pairwise R-FCN
- Co-localization with Category-Consistent Features and Geodesic Distance Propagation
- Video Representation Learning Using Discriminative Pooling
- Weakly Supervised Object Detection with Segmentation Collaboration
- Comparison Network for One-Shot Conditional Object Detection
- Utilizing the Instability in Weakly Supervised Object Detection
- Open-Vocabulary Object Detection Using Captions
- Saliency Guided End-to-End Learning for Weakly Supervised Object Detection
- Re-Identification with Consistent Attentive Siamese Networks
- Training object class detectors with click supervision
- Towards Noise-resistant Object Detection with Noisy Annotations
- Temporal Dynamic Graph LSTM for Action-driven Video Object Detection
- Visual Relationship Detection using Scene Graphs: A Survey
- Weakly Supervised Object Localization Using Size Estimates
- Progressive Object Transfer Detection
- Scalable Deep Learning Logo Detection
- Proposal Learning for Semi-Supervised Object Detection
- Collaborative Learning for Weakly Supervised Object Detection
- PCAMs: Weakly Supervised Semantic Segmentation Using Point Supervision
- Weakly Supervised Object Localization Using Things and Stuff Transfer
- Zooming into Face Forensics: A Pixel-level Analysis
- Optimizing Region Selection for Weakly Supervised Object Detection
- Semi-supervised reward learning for offline reinforcement learning
- Spatial Likelihood Voting with Self-Knowledge Distillation for Weakly Supervised Object Detection
- Towards Precise End-to-end Weakly Supervised Object Detection Network
- Object-Aware Instance Labeling for Weakly Supervised Object Detection
- Weakly Supervised Person Re-ID: Differentiable Graphical Learning and A New Benchmark
- Learning the semantic structure of objects from Web supervision
- MINI-Net: Multiple Instance Ranking Network for Video Highlight Detection
- Exploiting Playbacks in Unsupervised Domain Adaptation for 3D Object Detection
- Mixed Supervised Object Detection with Robust Objectness Transfer
- Exploiting Web Images for Weakly Supervised Object Detection
- Discover and Learn New Objects from Documentaries
- Dual-attention Focused Module for Weakly Supervised Object Localization
- Activity Driven Weakly Supervised Object Detection
- Spatio-Temporal Instance Learning: Action Tubes from Class Supervision
- Active learning with version spaces for object detection
- LID 2020: The Learning from Imperfect Data Challenge Results
- Visual Semantic Information Pursuit: A Survey
- Learning to Count Objects with Few Exemplar Annotations
- Audio Event and Scene Recognition: A Unified Approach using Strongly and Weakly Labeled Data
- Exploring Bottom-up and Top-down Cues with Attentive Learning for Webly Supervised Object Detection
- Weakly Supervised PET Tumor Detection Using Class Response
- Diverse Sampling for Self-Supervised Learning of Semantic Segmentation
- Image Captioning with Unseen Objects
- Learning from Counting: Leveraging Temporal Classification for Weakly Supervised Object Localization and Detection
- SLV: Spatial Likelihood Voting for Weakly Supervised Object Detection
- Modularized Textual Grounding for Counterfactual Resilience
- Caption Generation on Scenes with Seen and Unseen Object Categories
- Action Representation Using Classifier Decision Boundaries
- Weakly Supervised Arrhythmia Detection Based on Deep Convolutional Neural Network
- Discovery-and-Selection: Towards Optimal Multiple Instance Learning for Weakly Supervised Object Detection
- Towards Robust Pattern Recognition: A Review
- W-TALC: Weakly-supervised Temporal Activity Localization and Classification
- Multi-scale discriminative Region Discovery for Weakly-Supervised Object Localization
- Adaptively Denoising Proposal Collection for Weakly Supervised Object Localization
- Unsupervised Object Detection with LiDAR Clues
- Progressive Representation Adaptation for Weakly Supervised Object Localization
- Weakly-Supervised Multi-Person Action Recognition in 360 Videos
- Dual-attention Guided Dropblock Module for Weakly Supervised Object Localization
- EHSOD: CAM-Guided End-to-end Hybrid-Supervised Object Detection with Cascade Refinement
- Min-Entropy Latent Model for Weakly Supervised Object Detection
- Privacy-Aware Activity Classification from First Person Office Videos
- Weakly Supervised Object Localization and Detection: A Survey
- TS2C: Tight Box Mining with Surrounding Segmentation Context for Weakly Supervised Object Detection
- Generative Adversarial Networks for Unsupervised Object Co-localization
- Self-Guided Multiple Instance Learning for Weakly Supervised Disease Classification and Localization in Chest Radiographs
- MakeupBag: Disentangling Makeup Extraction and Application
- A Comparative Study on Effects of Original and Pseudo Labels for Weakly Supervised Learning for Car Localization Problem
- Learning Rich Representations For Structured Visual Prediction Tasks
- Spatio-Temporal Action Localization in a Weakly Supervised Setting
- Towards Accurate Localization by Instance Search
- Discriminative Video Representation Learning Using Support Vector Classifiers
- Anomaly Detection with Inexact Labels
- Self-learning Scene-specific Pedestrian Detectors using a Progressive Latent Model
- Two-phase weakly supervised object detection with pseudo ground truth mining