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

By Chenchen Zhu, Yihui He, Marios Savvides. 2019

917 citations recorded by the source corpus.

We motivate and present feature selective anchor-free (FSAF) module, a simple and effective building block for single-shot object detectors. It can be plugged into single-shot detectors with feature pyramid structure. The FSAF module addresses two limitations brought up by the conventional anchor-based detection: 1) heuristic-guided feature selection; 2) overlap-based anchor sampling. The general concept of the FSAF module is online feature selection applied to the training of multi-level anchor-free branches. Specifically, an anchor-free branch is attached to each level of the feature pyramid, allowing box encoding and decoding in the anchor-free manner at an arbitrary level. During training, we dynamically assign each instance to the most suitable feature level. At the time of inference, the FSAF module can work independently or jointly with anchor-based branches. We instantiate this concept with simple implementations of anchor-free branches and online feature selection strategy. Experimental results on the COCO detection track show that our FSAF module performs better than anchor-based counterparts while being faster. When working jointly with anchor-based branches, the FSAF module robustly improves the baseline RetinaNet by a large margin under various settings, while introducing nearly free inference overhead. And the resulting best model can achieve a state-of-the-art 44.6% mAP, outperforming all existing single-shot detectors on COCO.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

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

Bridging the Gap Between Anchor-Based and Anchor-Free Detection via Adaptive Training Sample Selection (2019, background): “FSAF [72] attaches an anchor-free branch with online feature selection to RetinaNet.”
Objects as Points (2019, background): “Nonetheless, over the past ﬁve years [19], this idea has achieved good empirical success [12,21,25,26,31,35,47,48,56,62,63].”
FCOS: Fully Convolutional One-Stage Object Detection (2019, neutral): “9 FSAF [34] ResNeXt-64x4d-101-FPN 42.”
MMDetection: Open MMLab Detection Toolbox and Benchmark (2019, background): “• FSAF [39]: a feature selective anchor-free module for single-stage detectors, proposed in 2019.”
YOLOv4: Optimal Speed and Accuracy of Object Detection (2020, neutral): “Feature Selective Anchor-Free Module for Single-Shot Object Detection [102] AB+FSAF ResNet-101 800 5.”
Conditional DETR for Fast Training Convergence (2021, methodology): “Typical methods include YOLOv1 [30], CornerNet [21], ExtremeNet [50], CenterNet [49, 6], FCOS [39] and others [23, 28, 52, 19, 51, 22, 15, 46, 47].”
SpatialFlow: Bridging All Tasks for Panoptic Segmentation (2019, background): “Recently, researchers focus on improving the localization quality of one-stage detectors and propose anchor-free algorithms [13], [52], [53], [54], [55].”
Probabilistic Anchor Assignment with IoU Prediction for Object Detection (2020, background): “have been introduced. These models predict various key points such as corners [18], extreme points [44], center points [6,33] or arbitrary feature points [38] induced from deformable convolution [5]. [45] combines anchor-based detectors with anchor-free detection by adding additional anchor-free reg”
Tinier-YOLO: A Real-Time Object Detection Method for Constrained Environments (2020, background): “Recent work of anchor-free one-stage detectors like FSAF [18] and CenterNet [19] achieved a better trade-off between speed and accuracy.”
Soft Anchor-Point Object Detection (2019, methodology): “FSAF [35], Guided Anchoring [26], FCOS [25] and FoveaBox [10] encode and decode the bounding boxes as anchor points and point-to-boundary distances.”
UAV-YOLOv8: A Small-Object-Detection Model Based on Improved YOLOv8 for UAV Aerial Photography Scenarios (2023, methodology): “However, in the context of visual tasks where small objects are the Both RetinaNet and FSAF use multiscale feature fusion to consider objects at different scales. primary detection objects, these two model structures cannot address the detection needs.”
VisDrone-DET2019: The Vision Meets Drone Object Detection in Image Challenge Results (2019, methodology): “2) We add feature selected anchor-free head (FSAF) [47] into RetinaNet, which improves the performance signiﬁcantly.”
IoU-uniform R-CNN: Breaking Through the Limitations of RPN (2019, methodology): “Instead of using anchor boxes, they predict bounding boxes in a per-pixel prediction fashion [33, 35] or keypoint-based fashion [18, 7].”
Group Fisher Pruning for Practical Network Compression (2021, methodology): “We prune one-stage meth-Group ods including RetinaNet (Lin et al., 2017b), FSAF (Zhu et al., 2019), ATSS (Zhang et al., 2020) and PAA (Kim & Lee, 2020), and two-stage method Faster R-CNN (Ren et al., 2015) to extensively validate the effectiveness of our method.”
Rethinking Transformer-based Set Prediction for Object Detection (2020, methodology): “We compare TSP-RCNN with multiple one-stage and two-stage object detection models [26, 31, 1, 28, 19, 40, 30, 3, 16, 35, 38] that also use ResNet-101 backbone or its deformable convolution network (DCN) [41] variant in Table 4.”