Abstract:There are still two problems in SDD causing some inaccurate results: (1) In the process of feature extraction, with the layer-by-layer acquisition of semantic information, local information is gradually lost, resulting into less representative feature maps; (2) During the Non-Maximum Suppression (NMS) algorithm due to inconsistency in classification and regression tasks, the classification confidence and predicted detection position cannot accurately indicate the position of the prediction boxes. Methods: In order to address these aforementioned issues, we propose a new architecture, a modified version of Single Shot Multibox Detector (SSD), named Precise Single Stage Detector (PSSD). Firstly, we improve the features by adding extra layers to SSD. Secondly, we construct a simple and effective feature enhancement module to expand the receptive field step by step for each layer and enhance its local and semantic information. Finally, we design a more efficient loss function to predict the IOU between the prediction boxes and ground truth boxes, and the threshold IOU guides classification training and attenuates the scores, which are used by the NMS algorithm. Main Results: Benefiting from the above optimization, the proposed model PSSD achieves exciting performance in real-time. Specifically, with the hardware of Titan Xp and the input size of 320 pix, PSSD achieves 33.8 mAP at 45 FPS speed on MS COCO benchmark and 81.28 mAP at 66 FPS speed on Pascal VOC 2007 outperforming state-of-the-art object detection models. Besides, the proposed model performs significantly well with larger input size. Under 512 pix, PSSD can obtain 37.2 mAP with 27 FPS on MS COCO and 82.82 mAP with 40 FPS on Pascal VOC 2007. The experiment results prove that the proposed model has a better trade-off between speed and accuracy.

Multi-scales Feature Integration Single Shot Multi-Box Detector on Small Object Detection

Small Object Detection Algorithm Based on Feature Pyramid-Enhanced Fusion SSD

Scale Pyramid Attention for Single Shot MultiBox Detector

FSSD: Feature Fusion Single Shot Multibox Detector

SSD: Single Shot MultiBox Detector

Detecting Small Objects in Thermal Images Using Single-Shot Detector

MSSD: multi-scale object detector based on spatial pyramid depthwise convolution and efficient channel attention mechanism

TDFSSD: Top-Down Feature Fusion Single Shot MultiBox Detector

Precise Single-stage Detector

Scale-Balanced Real-Time Object Detection with Varying Input-Image Resolution

Inception Single Shot MultiBox Detector for Object Detection

A Balanced Feature Fusion SSD for Object Detection

Extend the Shallow Part of Single Shot MultiBox Detector Via Convolutional Neural Network

MDSSD: Multi-scale Deconvolutional Single Shot Detector for Small Objects

Single-Shot Object Detection via Feature Enhancement and Channel Attention

Object Detection Based on the Improved Single Shot MultiBox Detector

FFR-SSD: feature fusion and reconstruction single shot detector for multi-scale object detection

Object Detection Using Deep Learning: Single Shot Detector with a Refined Feature-fusion Structure

Improvements Based on Feature Fusion Single Shot Multibox Detector

Efficient Featurized Image Pyramid Network for Single Shot Detector

Attention Based Single Shot Multibox Detector