FCOS: A Simple and Strong Anchor-Free Object Detector

By Zhi Tian, Chunhua Shen, Hao Chen, Tong He. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2020

995 citations recorded by the source corpus.

In computer vision, object detection is one of most important tasks, which underpins a few instance-level recognition tasks and many downstream applications. Recently one-stage methods have gained much attention over two-stage approaches due to their simpler design and competitive performance. Here we propose a fully convolutional one-stage object detector (FCOS) to solve object detection in a per-pixel prediction fashion, analogue to other dense prediction problems such as semantic segmentation. Almost all state-of-the-art object detectors such as RetinaNet, SSD, YOLOv3, and Faster R-CNN rely on pre-defined anchor boxes. In contrast, our proposed detector FCOS is anchor box free, as well as proposal free. By eliminating the pre-defined set of anchor boxes, FCOS completely avoids the complicated computation related to anchor boxes such as calculating the intersection over union (IoU) scores during training. More importantly, we also avoid all hyper-parameters related to anchor boxes, which are often sensitive to the final detection performance. With the only post-processing non-maximum suppression (NMS), we demonstrate a much simpler and flexible detection framework achieving improved detection accuracy. We hope that the proposed FCOS framework can serve as a simple and strong alternative for many other instance-level tasks. Code is available at: <inline-formula><tex-math notation="LaTeX">$\tt git.io/AdelaiDet$</tex-math><alternatives><mml:math><mml:mrow><mml:mi mathvariant="monospace">git</mml:mi><mml:mo>.</mml:mo><mml:mi mathvariant="monospace">io</mml:mi><mml:mo>/</mml:mo><mml:mi mathvariant="monospace">AdelaiDet</mml:mi></mml:mrow></mml:math><inline-graphic xlink:href="shen-ieq1-3032166.gif"/></alternatives></inline-formula>

Subfields: Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/554363727a07d39b23ad5b4e32f556685afd332a

How researchers see this paper

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

In Computer Vision and Pattern Recognition

Citations in the corpus

Towards Large-Scale Small Object Detection: Survey and Benchmarks (2022, methodology): “Information Loss: Current prevailing object detectors [1], [2], [3], [4], [5], [9] usually include a backbone network and a detection head, where the latter makes decision depends on the representation output by the former.”
VinVL: Revisiting Visual Representations in Vision-Language Models (2021, background): “OpenImages and Objects 365), 2) gaining new insights in object detection algorithms such as feature pyramid network [22], one-stage dense prediction [23], and anchor-free detectors [36], and 3) leveraging more powerful GPUs for training bigger models.”
YOLOv9: Learning What You Want to Learn Using Programmable Gradient Information (2024, methodology): “In terms of feature integration, improved PAN [37] or FPN [35] is often used as a tool, and then improved YOLOv3 head [49] or FCOS head [57, 58] is used as prediction head.”
YOLOv7: Trainable Bag-of-Freebies Sets New State-of-the-Art for Real-Time Object Detectors (2022, methodology): “Being able to become a stateof-the-art real-time object detector usually requires the following characteristics: (1) a faster and stronger network architecture; (2) a more effective feature integration method [8, 21, 29, 36, 43, 56, 71, 94]; (3) a more accurate detection method [66, 73, 74]; (4) a m”
Transformer-Based Visual Segmentation: A Survey (2023, background): “• Neck: Feature pyramid network (FPN) has been shown effective in object detection and instance segmentation [116], [117], [118] for scale variation modeling.”
Instances as Queries (2021, methodology): “Sparse R-CNN [42] build a query based detector on top of R-CNN architecture, while OneNet [40] and DeFCN [50] are end-to-end detector built upon the onestage FCOS [45].”
YOLOv10: Real-Time End-to-End Object Detection (2024, background): “Over the past years, substantial efforts have been directed towards developing efficient detectors [19, 57, 48, 34, 79, 75, 32, 31, 41].”
What's Behind the Couch? Directed Ray Distance Functions (DRDF) for 3D Scene Reconstruction (2021, neutral): “In many cases, the expected value is precisely what is desired like in object detection [59, 71] or in ARIMA models [44], weather prediction [66] but in others it leads to poor outcomes.”
Probabilistic two-stage detection (2021, methodology): “During testing, we set the short edge at 512 and the long edge up to 736 (Tian et al., 2020).”
ResT: An Efficient Transformer for Visual Recognition (2021, background): “Deep learning backbone architectures have been evolved for years and boost the performance of computer vision tasks such as classiﬁcation [5, 25, 32, 10], object detection [2, 39, 17, 24], and instance segmentation [9, 23, 30], etc.”
MiniVLM: A Smaller and Faster Vision-Language Model (2020, background): “Region feature is built on top of the object detector, and the detector can be two-stage [38, 8, 24, 55] or one-stage [37, 48, 29, 27, 6, 58, 53, 49].”
VinVL: Making Visual Representations Matter in Vision-Language Models (2021, background): “OpenImages and Objects 365), 2) gaining new insights in object detection algorithms such as feature pyramid network [23], one-stage dense prediction [24], and anchor-free detectors [37], and 3) leveraging more powerful GPUs for training bigger models.”