Learning Spatial Fusion for Single-Shot Object Detection

By Songtao Liu, Di Huang, Yunhong Wang. arXiv.org, 2019

863 citations recorded by the source corpus.

Pyramidal feature representation is the common practice to address the challenge of scale variation in object detection. However, the inconsistency across different feature scales is a primary limitation for the single-shot detectors based on feature pyramid. In this work, we propose a novel and data driven strategy for pyramidal feature fusion, referred to as adaptively spatial feature fusion (ASFF). It learns the way to spatially filter conflictive information to suppress the inconsistency, thus improving the scale-invariance of features, and introduces nearly free inference overhead. With the ASFF strategy and a solid baseline of YOLOv3, we achieve the best speed-accuracy trade-off on the MS COCO dataset, reporting 38.1% AP at 60 FPS, 42.4% AP at 45 FPS and 43.9% AP at 29 FPS. The code is available at this https URL

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Radiology, Nuclear Medicine and Imaging

View Paper: https://www.semanticscholar.org/paper/5a90e23b8c85d909d9589f45c46ed0f091fbf975

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Radiology, Nuclear Medicine and Imaging

Citations in the corpus

PP-YOLO: An Effective and Efficient Implementation of Object Detector (2020, neutral): “3% YOLOv3 + ASFF* [26] Darknet-53 800 29.”
AFPN: Asymptotic Feature Pyramid Network for Object Detection (2023, methodology): “We leverage ASFF [20] to assign varying spatial weights to the different levels of features during the multi-level feature fusion process, enhancing the signiﬁcance of pivotal levels and mitigate the impact of contradictory information from different objects.”
Track to Detect and Segment: An Online Multi-Object Tracker (2021, methodology): “Similar to [24], w is predicted by two convolution layers followed by softmax function.”
FFCA-YOLO for Small Object Detection in Remote Sensing Images (2024, background): “Since then, PANet [26], NAS-FPN [27], ASFF [28], and BiFPN [29] are proposed and achieve good results in object detection tasks.”
Towards Large-Scale Small Object Detection: Survey and Benchmarks (2022, background): “Noting that the gradient inconsistency encountered in FPN-based approaches deteriorates the representation ability of low-level features [92], SSPNet [93] highlights the features of speciﬁc scalesatdifferentlayersandemploystherelationshipofadjacent layers in FPN to accomplish proper feature sharing.”
PP-YOLOE: An evolved version of YOLO (2022, neutral): “6% YOLOv3 + ASFF* [17] Darknet-53 416 54 - 40.”
Effective Fusion Factor in FPN for Tiny Object Detection (2020, background): “ASFF [18] predicts the weight of features from different stages via a self-adaptive mechanism when fused again.”
RetinaNet With Difference Channel Attention and Adaptively Spatial Feature Fusion for Steel Surface Defect Detection (2020, methodology): “To solve the above-mentioned problems, this study proposes a novel RetinaNet with difference channel attention and adaptively spatial feature fusion (DEA_RetinaNet) [26] for steel surface defect detection.”
OTA: Optimal Transport Assignment for Object Detection (2021, methodology): “IoU based label assignment is proved effective and soon been adopted by many Faster R-CNN’s variants like [2, 12, 20, 42, 26, 49, 37], as well as many one-stage detectors like [31, 32, 25, 27, 23, 21].”