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.

Cited by in corpus (34)

• CenterNet: Keypoint Triplets for Object Detection
• IPOD: Intensive Point-based Object Detector for Point Cloud
• RetinaMask: Learning to predict masks improves state-of-the-art single-shot detection for free
• Single-Shot Refinement Neural Network for Object Detection
• Bottom-up Object Detection by Grouping Extreme and Center Points
• Online PCB Defect Detector On A New PCB Defect Dataset
• Augmentation for small object detection
• Data Augmentation for Object Detection via Progressive and Selective Instance-Switching
• Cascade RetinaNet: Maintaining Consistency for Single-Stage Object Detection
• GFD-SSD: Gated Fusion Double SSD for Multispectral Pedestrian Detection
• Residual Features and Unified Prediction Network for Single Stage Detection
• Consistent Optimization for Single-Shot Object Detection
• ChainerCV: a Library for Deep Learning in Computer Vision
• Extend the shallow part of Single Shot MultiBox Detector via Convolutional Neural Network
• Learning Efficient Detector with Semi-supervised Adaptive Distillation
• Feature Intertwiner for Object Detection
• Single Pixel Reconstruction for One-stage Instance Segmentation
• ALFA: Agglomerative Late Fusion Algorithm for Object Detection
• Improve Object Detection by Data Enhancement based on Generative Adversarial Nets
• Weaving Multi-scale Context for Single Shot Detector
• TACNet: Transition-Aware Context Network for Spatio-Temporal Action Detection
• A unified neural network for object detection, multiple object tracking and vehicle re-identification
• Rank of Experts: Detection Network Ensemble
• Detecting retail products in situ using CNN without human effort labeling
• DuBox: No-Prior Box Objection Detection via Residual Dual Scale Detectors
• Low-Power Computer Vision: Status, Challenges, Opportunities
• Towards Adversarially Robust Object Detection
• IMP: Instance Mask Projection for High Accuracy Semantic Segmentation of Things
• Multiple receptive fields and small-object-focusing weakly-supervised segmentation network for fast object detection
• DisplaceNet: Recognising Displaced People from Images by Exploiting Dominance Level
• FHEDN: A based on context modeling Feature Hierarchy Encoder-Decoder Network for face detection
• Object Detection in Specific Traffic Scenes using YOLOv2
• IvaNet: Learning to jointly detect and segment objets with the help of Local Top-Down Modules
• Detecting Lesion Bounding Ellipses With Gaussian Proposal Networks