FSSD: Feature Fusion Single Shot Multibox Detector

By Zuoxin Li, Fuqiang Zhou. arXiv.org, 2017

572 citations recorded by the source corpus.

SSD (Single Shot Multibox Detector) is one of the best object detection algorithms with both high accuracy and fast speed. However, SSD's feature pyramid detection method makes it hard to fuse the features from different scales. In this paper, we proposed FSSD (Feature Fusion Single Shot Multibox Detector), an enhanced SSD with a novel and lightweight feature fusion module which can improve the performance significantly over SSD with just a little speed drop. In the feature fusion module, features from different layers with different scales are concatenated together, followed by some down-sampling blocks to generate new feature pyramid, which will be fed to multibox detectors to predict the final detection results. On the Pascal VOC 2007 test, our network can achieve 82.7 mAP (mean average precision) at the speed of 65.8 FPS (frame per second) with the input size 300$\times$300 using a single Nvidia 1080Ti GPU. In addition, our result on COCO is also better than the conventional SSD with a large margin. Our FSSD outperforms a lot of state-of-the-art object detection algorithms in both aspects of accuracy and speed. Code is available at this https URL.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/115e8a4f76a57b893c985ea4f7530b90d071679c

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

Citations in the corpus

Deep Learning Sensor Fusion for Autonomous Vehicle Perception and Localization: A Review (2020, unknown): “In most cases, the improvement of accuracy is traded with the speed of detection [111].”
Recent Advances in Deep Learning for Object Detection (2019, background): “Thus, handling feature were later developed [109, 110, 109, 111, 112, 92, 113, 114, 115, 116, 117, 118, 119], with modiﬁcations to the feature pyramid block (see Fig.”
Tinier-YOLO: A Real-Time Object Detection Method for Constrained Environments (2020, background): “YOLO-V2 [8], YOLO-V3 [9] and the series of Single Shot Multi-box Detector (SSD) [13]–[16] have tried to solve the problem.”
MetaAnchor: Learning to Detect Objects with Customized Anchors (2018, unknown): ““default boxes” in [25], “priors” in [39] or “grid cells” in [30]) are employed in most of the state-of-the-art detection systems [39, 32, 22, 23, 25, 7, 11, 2, 31, 21, 35, 15].”
FAIR1M: A Benchmark Dataset for Fine-grained Object Recognition in High-Resolution Remote Sensing Imagery (2021, background): “The above large-scale datasets promote the development of algorithms on general object detection in natural images [25, 23, 47, 30, 62, 42, 40].”
KGSNet: Key-Point-Guided Super-Resolution Network for Pedestrian Detection in the Wild (2020, background): “The one-stage framework directly predicts object categories and regresses object bounding box based on the inputted image regions or anchors, where the famous works are [51], [52] and some follow-up works [53]–[57].”
Small-Object Detection in Remote Sensing Images with End-to-End Edge-Enhanced GAN and Object Detector Network (2020, methodology): “In [54], feature fusion SSD (FSSD) was proposed where features from different layers with different scales were concatenated together, and then some downsampling blocks were used to generate new feature pyramids.”
Multispectral Fusion for Object Detection with Cyclic Fuse-and-Refine Blocks (2020, unknown): “As implemented in SSD [13] and FSSD [14], a few data augmentation methods are applied, such as image random cropping, padding, ﬂipping and distorting for both visible and thermal images.”
A Survey of Deep Learning-Based Object Detection (2019, unknown): “Available: https://arxiv.org/abs/1901.03278 [110] X. Zhou, J. Zhuo, and P. Krähenbühl, ‘‘Bottom-up object detection by grouping extreme and center points,’’ 2019, arXiv:1901.08043.”
Multi-Scale Boosted Dehazing Network With Dense Feature Fusion (2020, methodology): “In [25, 37], the features at different scales are projected and concatenated by using the strided convolutional layer.”
Towards Adversarially Robust Object Detection (2019, methodology): “Experiments with different detector architectures (Receptive Field Block-based Detector (RFB) [28], Feature Fusion Single Shot Detector (FSSD) [24] and YOLO-V3 [40, 41]) and backbones (VGG16 [45], ResNet50 [19], DarkNet53 [39]) are also conducted for comprehensive evaluations.”