R-FCN-3000 at 30fps: Decoupling Detection and Classification

By Bharat Singh, Hengduo Li, Abhishek Sharma, Larry S. Davis. 2017

98 citations recorded by the source corpus.

We propose a modular approach towards large-scale real-time object detection by decoupling objectness detection and classification. We exploit the fact that many object classes are visually similar and share parts. Thus, a universal objectness detector can be learned for class-agnostic object detection followed by fine-grained classification using a (non)linear classifier. Our approach is a modification of the R-FCN architecture to learn shared filters for performing localization across different object classes. We trained a detector for 3000 object classes, called R-FCN-3000, that obtains an mAP of 34.9% on the ImageNet detection dataset. It outperforms YOLO-9000 by 18% while processing 30 images per second. We also show that the objectness learned by R-FCN-3000 generalizes to novel classes and the performance increases with the number of training object classes - supporting the hypothesis that it is possible to learn a universal objectness detector. Code will be made available.

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

RegionCLIP: Region-based Language-Image Pretraining (2021, methodology): “Inspired by [47, 64], we fused RPN objectness scores and category confidence scores by geometry mean.”
Learning Open-World Object Proposals without Learning to Classify (2021, background): “Approaches under semisupervised setup [29], [30], [31], [32] employ the weak imagelevel labels for novel classes as well as box-level labels for base classes.”
Object Detection in Optical Remote Sensing Images: A Survey and A New Benchmark (2019, background): “Also, Singh et al. proposed a novel detector, called R ‐ FCN ‐ 3000 (Singh et al., 2018a), towards large ‐ scale real ‐ time object detection for 3000 object classes.”
Soft Sampling for Robust Object Detection (2018, unknown): “As deep learning provided a large boost in performance to visual recognition problems, detectors like YOLO-9000 [17] and R-FCN-3000 [23] leveraged convolutional neural networks [12] for efficiently detecting a large number of classes.”
Learning From Noisy Anchors for One-Stage Object Detection (2019, background): “A multitude of detectors have been newly proposed based on these frameworks [32, 37, 25, 2, 44, 7, 6, 46, 4].”
What leads to generalization of object proposals? (2020, background): “Past works like RFCNN3000 [30] argued that a single super class might be sufficient to obtain good”
Detecting 11K Classes: Large Scale Object Detection Without Fine-Grained Bounding Boxes (2019, methodology): “One is the 3 K set with bounding box annotations, similar to what is used in [27].”
DC-SPP-YOLO: Dense Connection and Spatial Pyramid Pooling Based YOLO for Object Detection (2019, methodology): “In the past two years, based on the Faster R-CNN and RFCN, RRPN (Rotation Region Proposal Networks) [12], R-FCN-3000 [13] and other object proposal-based approaches [14] [15] of which the detection accuracy was further improved were presented.”
Exploring Object Relation in Mean Teacher for Cross-Domain Detection (2019, background): “Next, a few subsequent works [7, 8, 18, 22, 23, 33, 46] strive to improve the accuracy and speed of two-stage detectors.”
Detecting Twenty-thousand Classes using Image-level Supervision (2022, unknown): “Large-vocabulary object detection [18, 45, 53, 69] requires detecting 1000+ classes.”
Cross-Domain Adaptive Teacher for Object Detection (2021, background): “Afterwards, several anchor-based works [5,6,15,20,26,37] are proposed to improve the performance and efficacy.”
Forest R-CNN: Large-Vocabulary Long-Tailed Object Detection and Instance Segmentation (2020, background): “Recently, [29, 31, 35, 36] also strive to solve the problem of object recognition in a large vocabulary.”