Feature Selective Anchor-Free Module for Single-Shot Object Detection

Zhu, Chenchen; He, Yihui; Savvides, Marios

doi:10.1109/cvpr.2019.00093

Cited by 823 publications

(410 citation statements)

References 45 publications

Supporting

Mentioning

408

Contrasting

Unclassified

Order By: Relevance

“…Besides of focal loss, gradient harmonized mechanism (GHM) [11] is introduced to reduce the impacts of easy negatives for effect learning. Anchor-free module [30] is proposed for box encoding and decoding at arbitrary level in feature pyramid, as well as online feature selection to improve the single-stage accuracy.…”

Section: A Object Detectionmentioning

confidence: 99%

HAR-Net: Joint Learning of Hybrid Attention for Single-Stage Object Detection

Wang

2020

IEEE Trans. on Image Process.

View full text Add to dashboard Cite

Object detection has been a challenging task in computer vision. Although significant progress has been made in object detection with deep neural networks, the attention mechanism is far from development. In this paper, we propose the hybrid attention mechanism for single-stage object detection. First, we present the modules of spatial attention, channel attention and aligned attention for single-stage object detection. In particular, stacked dilated convolution layers with symmetrically fixed rates are constructed to learn spatial attention. The channel attention is proposed with the cross-level group normalization and squeezeand-excitation module. Aligned attention is constructed with organized deformable filters. Second, the three kinds of attention are unified to construct the hybrid attention mechanism. We then embed the hybrid attention into Retina-Net and propose the efficient single-stage HAR-Net for object detection. The attention modules and the proposed HAR-Net are evaluated on the COCO detection dataset. Experiments demonstrate that hybrid attention can significantly improve the detection accuracy and the HAR-Net can achieve the state-of-the-art 45.8% mAP, outperform existing single-stage object detectors.

show abstract

Section: A Object Detectionmentioning

confidence: 99%

HAR-Net: Joint Learning of Hybrid Attention for Single-Stage Object Detection

Wang

2020

IEEE Trans. on Image Process.

View full text Add to dashboard Cite

show abstract

“…Since then, most of the two-stage detectors [8,28,29] and the one-stage detectors [10,11,30] (including ship detectors [14,16]) introduce anchor strategy, and produce promising detection performance. However, these anchor-based detectors may suffer from some limitations [31,32].…”

Section: Encoder-decoder Network For Paired Semantic Segmentationmentioning

confidence: 99%

Ship Detection under Complex Backgrounds Based on Accurate Rotated Anchor Boxes from Paired Semantic Segmentation

Xiao

Zhou

et al. 2019

Remote Sensing

View full text Add to dashboard Cite

It is still challenging to effectively detect ship objects in optical remote-sensing images with complex backgrounds. Many current CNN-based one-stage and two-stage detection methods usually first predefine a series of anchors with various scales, aspect ratios and angles, and then the detection results can be outputted by performing once or twice classification and bounding box regression for predefined anchors. However, most of the defined anchors have relatively low accuracy, and are useless for the following classification and regression. In addition, the preset anchors are not robust to produce good performance for other different detection datasets. To avoid the above problems, in this paper we design a paired semantic segmentation network to generate more accurate rotated anchors with smaller numbers. Specifically, the paired segmentation network predicts four parts (i.e., top-left, bottom-right, top-right, and bottom-left parts) of ships. By combining paired top-left and bottom-right parts (or top-right and bottom-left parts), we can take the minimum bounding box of these two parts as the rotated anchor. This way can be more robust to different ship datasets, and the generated anchors are more accurate and have fewer numbers. Furthermore, to effectively use fine-scale detail information and coarse-scale semantic information, we use the magnified convolutional features to classify and regress the generated rotated anchors. Meanwhile, the horizontal minimum bounding box of the rotated anchor is also used to combine more context information. We compare the proposed algorithm with state-of-the-art object-detection methods for natural images and ship-detection methods, and demonstrate the superiority of our method.

show abstract

“…Backbone mAP CornerNet [3] Hourglass-104 40.8 CornerNet (Multi-Scale) [3] Hourglass-104 42.1 RetinaNet [6] ResNeXt-101-FPN 40.8 FSAF [7] ResNeXt-101-FPN 42.3 FSAF (Multi-Scale) [7] ResNeXt 101-FPN 44.6 CenterNet [2] Hourglass-104 44.9 CenterNet (Multi-Scale) [2] Hourglass-104 47.0 KP-xNet ResNeXt-101-X 44.7 KP-xNet (Multi-Scale)…”

Section: Architecturementioning

confidence: 99%

“…Single-shot detectors can also be split into two categories; anchor based detectors [6,7] and key-point based [5], where there are different output layers are assigned at each scale. Note we do not show the skip connections for the sake of simplicity.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Matrix Nets: A New Deep Architecture for Object Detection

Rashwan

Kalra²,

Poupart

2019

2019 IEEE/CVF International Conference on Computer Vision Workshop (ICCVW)

View full text Add to dashboard Cite

Figure 1. Model size (millions of parameters) vs accuracy (average precision) reported on MSCOCO test-dev for single-shot detectors. Enabled by our MatrixNet backbone, our model outperforms all other single-shot architectures while being smaller in size. AbstractWe present Matrix Nets (xNets), a new deep architecture for object detection. xNets map objects with different sizes and aspect ratios into layers where the sizes and the aspect ratios of the objects within their layers are nearly uniform. Hence, xNets provide a scale and aspect ratio aware architecture. We leverage xNets to enhance key-points based object detection. Our architecture achieves mAP of 47.8 on MS COCO, which is higher than any other single-shot detector while using half the number of parameters and training 3x faster than the next best architecture.

show abstract

Feature Selective Anchor-Free Module for Single-Shot Object Detection

Cited by 823 publications

References 45 publications

HAR-Net: Joint Learning of Hybrid Attention for Single-Stage Object Detection

HAR-Net: Joint Learning of Hybrid Attention for Single-Stage Object Detection

Ship Detection under Complex Backgrounds Based on Accurate Rotated Anchor Boxes from Paired Semantic Segmentation

Matrix Nets: A New Deep Architecture for Object Detection

Contact Info

Product

Resources

About