DSSD : Deconvolutional Single Shot Detector

Abstract

The main contribution of this paper is an approach for introducing additional context into state-of-the-art general object detection. To achieve this we first combine a state-of-the-art classifier (Residual-101[14]) with a fast detection framework (SSD[18]). We then augment SSD+Residual-101 with deconvolution layers to introduce additional large-scale context in object detection and improve accuracy, especially for small objects, calling our resulting system DSSD for deconvolutional single shot detector. While these two contributions are easily described at a high-level, a naive implementation does not succeed. Instead we show that carefully adding additional stages of learned transformations, specifically a module for feed-forward connections in deconvolution and a new output module, enables this new approach and forms a potential way forward for further detection research. Results are shown on both PASCAL VOC and COCO detection. Our DSSD with $513\times 513$ input achieves 81.5% mAP on VOC2007 test, 80.0% mAP on VOC2012 test, and 33.2% mAP on COCO, outperforming a state-of-the-art method R-FCN[3] on each dataset.

References in corpus (1)

• HyperNet: Towards Accurate Region Proposal Generation and Joint Object Detection

Cited by in corpus (234)

• YOLOv3: An Incremental Improvement
• Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation
• Object Detection in Optical Remote Sensing Images: A Survey and A New Benchmark
• Focal Loss for Dense Object Detection
• A Survey of Deep Learning-based Object Detection
• Distance-IoU Loss: Faster and Better Learning for Bounding Box Regression
• FCOS: Fully Convolutional One-Stage Object Detection
• Towards Large-Scale Small Object Detection: Survey and Benchmarks
• Slim-neck by GSConv: A lightweight-design for real-time detector architectures
• Object Detection in 20 Years: A Survey
• Learning Spatial Fusion for Single-Shot Object Detection
• FSSD: Feature Fusion Single Shot Multibox Detector
• DC-SPP-YOLO: Dense Connection and Spatial Pyramid Pooling Based YOLO for Object Detection
• Path Aggregation Network for Instance Segmentation
• Centralized Feature Pyramid for Object Detection
• Deep Learning for Generic Object Detection: A Survey
• PP-YOLO: An Effective and Efficient Implementation of Object Detector
• CenterNet: Keypoint Triplets for Object Detection
• CornerNet-Lite: Efficient Keypoint Based Object Detection
• R3Det: Refined Single-Stage Detector with Feature Refinement for Rotating Object
• Speed/accuracy trade-offs for modern convolutional object detectors
• CornerNet: Detecting Objects as Paired Keypoints
• SCRDet: Towards More Robust Detection for Small, Cluttered and Rotated Objects
• RepPoints: Point Set Representation for Object Detection
• Few-shot Object Detection on Remote Sensing Images
• Real-Time Fruit Recognition and Grasping Estimation for Autonomous Apple Harvesting
• Towards Accurate One-Stage Object Detection with AP-Loss
• IPOD: Intensive Point-based Object Detector for Point Cloud
• Receptive Field Block Net for Accurate and Fast Object Detection
• RetinaMask: Learning to predict masks improves state-of-the-art single-shot detection for free
• Recent Advances in Object Detection in the Age of Deep Convolutional Neural Networks
• Single-Shot Refinement Neural Network for Object Detection
• Scene Text Detection and Recognition: The Deep Learning Era
• Bottom-up Object Detection by Grouping Extreme and Center Points
• Tiny-DSOD: Lightweight Object Detection for Resource-Restricted Usages
• Layer-specific Optimization for Mixed Data Flow with Mixed Precision in FPGA Design for CNN-based Object Detectors
• AP-Loss for Accurate One-Stage Object Detection
• Scale-Aware Trident Networks for Object Detection
• Object Detection with Deep Learning: A Review
• Enhancing Geometric Factors in Model Learning and Inference for Object Detection and Instance Segmentation
• Online PCB Defect Detector On A New PCB Defect Dataset
• Attention-guided Context Feature Pyramid Network for Object Detection
• Gradient Harmonized Single-stage Detector
• Augmentation for small object detection
• Face Detection Using Improved Faster RCNN
• MDSSD: Multi-scale Deconvolutional Single Shot Detector for Small Objects
• Revisiting Feature Alignment for One-stage Object Detection
• Learning Modulated Loss for Rotated Object Detection
• A Systematic IoU-Related Method: Beyond Simplified Regression for Better Localization
• Convolutional Neural Networks with Gated Recurrent Connections
• Data Augmentation for Object Detection via Progressive and Selective Instance-Switching
• Continuous Human Action Recognition for Human-Machine Interaction: A Review
• Resolving Class Imbalance in Object Detection with Weighted Cross Entropy Losses
• Detect Faces Efficiently: A Survey and Evaluations
• Cascade RetinaNet: Maintaining Consistency for Single-Stage Object Detection
• Global Aggregation then Local Distribution in Fully Convolutional Networks
• PDNet: Toward Better One-Stage Object Detection With Prediction Decoupling
• Distilling Object Detectors with Task Adaptive Regularization
• AugFPN: Improving Multi-scale Feature Learning for Object Detection
• Decoupled Classification Refinement: Hard False Positive Suppression for Object Detection
• ThunderNet: Towards Real-time Generic Object Detection
• Object Detection for Comics using Manga109 Annotations
• Multi-Oriented Scene Text Detection via Corner Localization and Region Segmentation
• Multi-adversarial Faster-RCNN for Unrestricted Object Detection
• Rethinking Classification and Localization for Object Detection
• Sequence Level Semantics Aggregation for Video Object Detection
• FCOS: A simple and strong anchor-free object detector
• GFD-SSD: Gated Fusion Double SSD for Multispectral Pedestrian Detection
• CFENet: An Accurate and Efficient Single-Shot Object Detector for Autonomous Driving
• Objects detection for remote sensing images based on polar coordinates
• VarifocalNet: An IoU-aware Dense Object Detector
• Exploring Object Relation in Mean Teacher for Cross-Domain Detection
• MonoGRNet: A Geometric Reasoning Network for Monocular 3D Object Localization
• TANet: Robust 3D Object Detection from Point Clouds with Triple Attention
• RefineDetLite: A Lightweight One-stage Object Detection Framework for CPU-only Devices
• A Survey of Modern Object Detection Literature using Deep Learning
• Improving Object Detection from Scratch via Gated Feature Reuse
• Recent Advances in Deep Learning for Object Detection
• Feature Agglomeration Networks for Single Stage Face Detection
• Computing Systems for Autonomous Driving: State-of-the-Art and Challenges
• Detection in Crowded Scenes: One Proposal, Multiple Predictions
• Residual Features and Unified Prediction Network for Single Stage Detection
• QueryDet: Cascaded Sparse Query for Accelerating High-Resolution Small Object Detection
• Consistent Optimization for Single-Shot Object Detection
• Multi-granularity Generator for Temporal Action Proposal
• Intrinsic Relationship Reasoning for Small Object Detection
• SFace: An Efficient Network for Face Detection in Large Scale Variations
• A survey of Object Classification and Detection based on 2D/3D data
• ChainerCV: a Library for Deep Learning in Computer Vision
• Visibility Guided NMS: Efficient Boosting of Amodal Object Detection in Crowded Traffic Scenes
• Extend the shallow part of Single Shot MultiBox Detector via Convolutional Neural Network
• IoU Loss for 2D/3D Object Detection
• Structured Knowledge Distillation for Dense Prediction
• Surround-view Fisheye BEV-Perception for Valet Parking: Dataset, Baseline and Distortion-insensitive Multi-task Framework
• Cross-dataset Training for Class Increasing Object Detection
• Target Driven Instance Detection
• Domain Adaptive Object Detection via Asymmetric Tri-way Faster-RCNN
• Hit-Detector: Hierarchical Trinity Architecture Search for Object Detection
• CenterFace: Joint Face Detection and Alignment Using Face as Point
• BorderDet: Border Feature for Dense Object Detection
• C-RPNs: Promoting Object Detection in real world via a Cascade Structure of Region Proposal Networks
• M2Det: A Single-Shot Object Detector based on Multi-Level Feature Pyramid Network
• Feature Intertwiner for Object Detection
• Single-Shot Object Detection with Enriched Semantics
• SM-NAS: Structural-to-Modular Neural Architecture Search for Object Detection
• Learning Efficient Detector with Semi-supervised Adaptive Distillation
• Interpolation-based semi-supervised learning for object detection
• Modeling Visual Context is Key to Augmenting Object Detection Datasets
• Deep Adaptive Proposal Network for Object Detection in Optical Remote Sensing Images
• Control Distance IoU and Control Distance IoU Loss Function for Better Bounding Box Regression
• MimicDet: Bridging the Gap Between One-Stage and Two-Stage Object Detection
• Multi-scale Location-aware Kernel Representation for Object Detection
• Single-Shot Bidirectional Pyramid Networks for High-Quality Object Detection
• AI Matrix: A Deep Learning Benchmark for Alibaba Data Centers
• Underwater object detection using Invert Multi-Class Adaboost with deep learning
• How To Extract Fashion Trends From Social Media? A Robust Object Detector With Support For Unsupervised Learning
• Single Pixel Reconstruction for One-stage Instance Segmentation
• A Survey on Deep Domain Adaptation and Tiny Object Detection Challenges, Techniques and Datasets
• Autonomous Robotic Drilling System for Mice Cranial Window Creation: An Evaluation with an Egg Model
• TKD: Temporal Knowledge Distillation for Active Perception
• PolarMask++: Enhanced Polar Representation for Single-Shot Instance Segmentation and Beyond
• ScratchDet: Training Single-Shot Object Detectors from Scratch
• MultiResolution Attention Extractor for Small Object Detection
• Efficient Coarse-to-Fine Non-Local Module for the Detection of Small Objects
• NETNet: Neighbor Erasing and Transferring Network for Better Single Shot Object Detection
• Weaving Multi-scale Context for Single Shot Detector
• Improve Object Detection by Data Enhancement based on Generative Adversarial Nets
• TACNet: Transition-Aware Context Network for Spatio-Temporal Action Detection
• ALFA: Agglomerative Late Fusion Algorithm for Object Detection
• LapNet : Automatic Balanced Loss and Optimal Assignment for Real-Time Dense Object Detection
• Leveraging Bottom-Up and Top-Down Attention for Few-Shot Object Detection
• OmniPD: One-Step Person Detection in Top-View Omnidirectional Indoor Scenes
• MDFN: Multi-Scale Deep Feature Learning Network for Object Detection
• PointAtrousGraph: Deep Hierarchical Encoder-Decoder with Point Atrous Convolution for Unorganized 3D Points
• On the Importance of Visual Context for Data Augmentation in Scene Understanding
• SGPN: Similarity Group Proposal Network for 3D Point Cloud Instance Segmentation
• Deep Regionlets for Object Detection
• IPG-Net: Image Pyramid Guidance Network for Small Object Detection
• Feature Selective Small Object Detection via Knowledge-based Recurrent Attentive Neural Network
• A Novel Automation-Assisted Cervical Cancer Reading Method Based on Convolutional Neural Network
• MBAPose: Mask and Bounding-Box Aware Pose Estimation of Surgical Instruments with Photorealistic Domain Randomization
• A unified neural network for object detection, multiple object tracking and vehicle re-identification
• Zoom Out-and-In Network with Map Attention Decision for Region Proposal and Object Detection
• Context-Aware Single-Shot Detector
• Exploring Multi-Branch and High-Level Semantic Networks for Improving Pedestrian Detection
• Detecting retail products in situ using CNN without human effort labeling
• Rank of Experts: Detection Network Ensemble
• Fire SSD: Wide Fire Modules based Single Shot Detector on Edge Device
• Pixel and Feature Level Based Domain Adaption for Object Detection in Autonomous Driving
• Deep Regionlets: Blended Representation and Deep Learning for Generic Object Detection
• Low-Power Computer Vision: Status, Challenges, Opportunities
• DuBox: No-Prior Box Objection Detection via Residual Dual Scale Detectors
• Pseudo-IoU: Improving Label Assignment in Anchor-Free Object Detection
• iffDetector: Inference-aware Feature Filtering for Object Detection
• Object Detection based on Region Decomposition and Assembly
• SaccadeNet: A Fast and Accurate Object Detector
• Parallel Residual Bi-Fusion Feature Pyramid Network for Accurate Single-Shot Object Detection
• 3D Aggregated Faster R-CNN for General Lesion Detection
• Collage Inference: Using Coded Redundancy for Low Variance Distributed Image Classification
• RAPiD: Rotation-Aware People Detection in Overhead Fisheye Images
• Enhancing a Neurocognitive Shared Visuomotor Model for Object Identification, Localization, and Grasping With Learning From Auxiliary Tasks
• COMET: Context-Aware IoU-Guided Network for Small Object Tracking
• Triply Supervised Decoder Networks for Joint Detection and Segmentation
• NOH-NMS: Improving Pedestrian Detection by Nearby Objects Hallucination
• Deep Feature Pyramid Reconfiguration for Object Detection
• Dually Supervised Feature Pyramid for Object Detection and Segmentation
• Multiple Anchor Learning for Visual Object Detection
• Towards Adversarially Robust Object Detection
• CPM R-CNN: Calibrating Point-guided Misalignment in Object Detection
• MDCN: Multi-Scale, Deep Inception Convolutional Neural Networks for Efficient Object Detection
• Crowdsourced-based Deep Convolutional Networks for Urban Flood Depth Mapping
• Developing a Compressed Object Detection Model based on YOLOv4 for Deployment on Embedded GPU Platform of Autonomous System
• Pedestrian Detection with Autoregressive Network Phases
• Optical Braille Recognition Using Object Detection CNN
• Propose-and-Attend Single Shot Detector
• Towards Better Object Detection in Scale Variation with Adaptive Feature Selection
• Object-aware Feature Aggregation for Video Object Detection
• Multiple receptive fields and small-object-focusing weakly-supervised segmentation network for fast object detection
• Probabilistic Ranking-Aware Ensembles for Enhanced Object Detections
• FaPN: Feature-aligned Pyramid Network for Dense Image Prediction
• IMP: Instance Mask Projection for High Accuracy Semantic Segmentation of Things
• Temporally Identity-Aware SSD with Attentional LSTM
• HR-RCNN: Hierarchical Relational Reasoning for Object Detection
• Image Captioning with Unseen Objects
• Semi-Anchored Detector for One-Stage Object Detection
• Reprojection R-CNN: A Fast and Accurate Object Detector for 360° Images
• A Single-shot Object Detector with Feature Aggragation and Enhancement
• DisplaceNet: Recognising Displaced People from Images by Exploiting Dominance Level
• TLGAN: document Text Localization using Generative Adversarial Nets
• Multi-Grid Redundant Bounding Box Annotation for Accurate Object Detection
• Interflow: Aggregating Multi-layer Feature Mappings with Attention Mechanism
• Small Object Detection Based on Modified FSSD and Model Compression
• Tackling Variabilities in Autonomous Driving
• Mask R-CNN Based Object Detection for Intelligent Wireless Power Transfer
• EnsNet: Ensconce Text in the Wild
• Dive Deeper Into Box for Object Detection
• ASSD: Attentive Single Shot Multibox Detector
• LIP: Learning Instance Propagation for Video Object Segmentation
• Analysis and a Solution of Momentarily Missed Detection for Anchor-based Object Detectors
• Scale-wise Convolution for Image Restoration
• FPAN: Fine-grained and Progressive Attention Localization Network for Data Retrieval
• Focal Loss Dense Detector for Vehicle Surveillance
• ProbaNet: Proposal-balanced Network for Object Detection
• PiaNet: A pyramid input augmented convolutional neural network for GGO detection in 3D lung CT scans
• Ventral-Dorsal Neural Networks: Object Detection via Selective Attention
• Robust 2D/3D Vehicle Parsing in CVIS
• An Adaptive Video Acquisition Scheme for Object Tracking and its Performance Optimization
• Trained Model Fusion for Object Detection using Gating Network
• Detecting Lesion Bounding Ellipses With Gaussian Proposal Networks
• Leveraging Orientation for Weakly Supervised Object Detection with Application to Firearm Localization
• A DCNN-based Arbitrarily-Oriented Object Detector for Quality Control and Inspection Application
• RMOPP: Robust Multi-Objective Post-Processing for Effective Object Detection
• IvaNet: Learning to jointly detect and segment objets with the help of Local Top-Down Modules
• Adaptive Remote Sensing Image Attribute Learning for Active Object Detection
• Fine-Grained Vehicle Perception via 3D Part-Guided Visual Data Augmentation
• FHEDN: A based on context modeling Feature Hierarchy Encoder-Decoder Network for face detection
• Dense Multiscale Feature Fusion Pyramid Networks for Object Detection in UAV-Captured Images
• ScarfNet: Multi-scale Features with Deeply Fused and Redistributed Semantics for Enhanced Object Detection
• PosNeg-Balanced Anchors with Aligned Features for Single-Shot Object Detection
• Revisiting the Loss Weight Adjustment in Object Detection
• Object Detection in Specific Traffic Scenes using YOLOv2
• Fast Object Detection with Latticed Multi-Scale Feature Fusion
• Densely Semantic Enhancement for Domain Adaptive Region-free Detectors
• Layer-wise Customized Weak Segmentation Block and AIoU Loss for Accurate Object Detection
• POD: Practical Object Detection with Scale-Sensitive Network
• Attention-based Neural Network for Driving Environment Complexity Perception
• Pixel-Semantic Revise of Position Learning A One-Stage Object Detector with A Shared Encoder-Decoder
• Non-local RoIs for Instance Segmentation
• Boundary Distribution Estimation for Precise Object Detection
• Automatic Visual Inspection of Rare Defects: A Framework based on GP-WGAN and Enhanced Faster R-CNN
• USB: Universal-Scale Object Detection Benchmark
• Feature Fusion Detector for Semantic Cognition of Remote Sensing
• Object Detection in the Context of Mobile Augmented Reality
• Non-local RoI for Cross-Object Perception