Enhancement of SSD by concatenating feature maps for object detection

Jeong, Jisoo; Park, Hyojin; Kwak, Nojun

doi:10.48550/arxiv.1705.09587

Cited by 69 publications

(83 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The main feature fusion methods in the field of deep learning are divided into two categories: multi-network feature fusion [3,25,14] and single network feature fusion [11,3]. Multi-network feature fusion is always accompanied by multi-scale features, like Inception Model [25] or a set of multiple independent networks [3].…”

Section: Feature Fusion Methodsmentioning

confidence: 99%

NeuroTreeNet: A New Method to Explore Horizontal Expansion Network

Lou,

Luo,

Fan

et al. 2018

Preprint

View full text Add to dashboard Cite

It is widely recognized that the deeper networks or networks with more feature maps have better performance. Existing studies mainly focus on extending the network depth and increasing the feature maps of networks. At the same time, horizontal expansion network (e.g. Inception Model) as an alternative way to improve network performance has not been fully investigated. Accordingly, we proposed Neu-roTreeNet (NTN), as a new horizontal extension network through the combination of random forest and Inception Model. Based on the tree structure, in which each branch represents a network and the root node features are shared to child nodes, network parameters are effectively reduced. By combining all features of leaf nodes, even less feature maps achieved better performance. In addition, the relationship between tree structure and the performance of NTN was investigated in depth. Comparing to other networks (e.g. VDSR 5) with equal magnitude parameters, our model showed preferable performance in super resolution reconstruction task.

show abstract

Section: Feature Fusion Methodsmentioning

confidence: 99%

NeuroTreeNet: A New Method to Explore Horizontal Expansion Network

Lou,

Luo,

Fan

et al. 2018

Preprint

View full text Add to dashboard Cite

show abstract

“…Small object detection Recently, several ideas has been proposed for detecting small object [10,2,7,8]. Liu et al [10] augmented small object data by reducing the size of large objects for overcoming the not-enough-data problem.…”

Section: Related Workmentioning

confidence: 99%

“…However, it has the limitation of increased model complexity and slow down an speed due to applying deconvolution module to all feature maps. R-SSD [7] combines features of different scales through pooling and deconvolution and obtained improved accuracy and speed compared to DSSD. Li et al [8] uses Generative Adversarial Network(GAN) [5] to generate high-resolution features using low-resolution features as input to GAN.…”

Section: Related Workmentioning

confidence: 99%

Small Object Detection using Context and Attention

Lim¹,

Astrid²,

Yoon³

et al. 2019

Preprint

View full text Add to dashboard Cite

There are many limitations applying object detection algorithm on various environments. Especially detecting small objects is still challenging because they have lowresolution and limited information. We propose an object detection method using context for improving accuracy of detecting small objects. The proposed method uses additional features from different layers as context by concatenating multi-scale features. We also propose object detection with attention mechanism which can focus on the object in image, and it can include contextual information from target layer. Experimental results shows that proposed method also has higher accuracy than conventional SSD on detecting small objects. Also, for 300×300 input, we achieved 78.1% Mean Average Precision (mAP) on the PAS-CAL VOC2007 test set.

show abstract

“…Although deconvolution approaches have been very successful in the task of extracting the semantic information of the target being detected [8,11,13], adding too much deconvolution layers will inevitably lead to a significant increase in computational time-consuming [8,11], while providing too little semantic information will certainly lose some detection performance [13]. Based on the above consideration, three extension modules are applied in our proposed framework as shown in Figure 1, which have extended right amount of semantic information and can get good enough results without losing too much detection speed.…”

Section: Extension Modulementioning

confidence: 99%

Extend the shallow part of single shot multibox detector via convolutional neural network

Zheng

Zhao

2018

Tenth International Conference on Digital Image Processing (ICDIP 2018)

View full text Add to dashboard Cite

Single Shot MultiBox Detector (SSD) is one of the fastest algorithms in the current object detection field, which uses fully convolutional neural network to detect all scaled objects in an image. Deconvolutional Single Shot Detector (DSSD) is an approach which introduces more context information by adding the deconvolution module to SSD. And the mean Average Precision (mAP) of DSSD on PASCAL VOC2007 is improved from SSD's 77.5% to 78.6%. Although DSSD obtains higher mAP than SSD by 1.1%, the frames per second (FPS) decreases from 46 to 11.8. In this paper, we propose a single stage end-to-end image detection model called ESSD to overcome this dilemma. Our solution to this problem is to cleverly extend better context information for the shallow layers of the best single stage (e.g. SSD) detectors. Experimental results show that our model can reach 79.4% mAP, which is higher than DSSD and SSD by 0.8 and 1.9 points respectively.Meanwhile, our testing speed is 25 FPS in Titan X GPU which is more than double the original DSSD.

show abstract

Enhancement of SSD by concatenating feature maps for object detection

Cited by 69 publications

References 0 publications

NeuroTreeNet: A New Method to Explore Horizontal Expansion Network

NeuroTreeNet: A New Method to Explore Horizontal Expansion Network

Small Object Detection using Context and Attention

Extend the shallow part of single shot multibox detector via convolutional neural network

Contact Info

Product

Resources

About