Region-Based Convolutional Networks for Accurate Object Detection and Segmentation

By Ross B. Girshick, Jeff Donahue, Trevor Darrell, Jitendra Malik. 2016

2,661 citations recorded by the source corpus.

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

1. Luc Van Gool
2. Trevor Darrell
3. Alan L. Yuille
4. Bernt Schiele
5. Dacheng Tao
6. Jiashi Feng
7. Xiaogang Wang
8. Li Fei-Fei

In Computer Vision and Pattern Recognition

1. Luc Van Gool
2. Dacheng Tao
3. Ming-Hsuan Yang
4. Xiaogang Wang
5. Yi Yang
6. Trevor Darrell
7. Jitendra Malik
8. Alan L. Yuille

Citations in the corpus

• Fully convolutional networks for semantic segmentation (2014, methodology): “s. Hariharan et al. [14] and Gupta et al. [15] likewise adapt deep classification nets to semantic segmentation, but do so in hybrid proposal-classifier models. These approaches finetune an R-CNN system [5] by sampling bounding boxes and/or region proposals for detection, semantic segmentation, and ins”
• Interpreting Deep Visual Representations via Network Dissection (2017, methodology): “patches that maximize activation of hidden individual units [1], [8], [9], or the global feature space level by using variants of backpropagation to identify or generate salient image features [10], [11].”
• Dilated Residual Networks (2017, background): “Image classification is most often a proxy task that is used to pretrain a model before it is transferred to other applications that involve more detailed scene understanding [4, 10].”
• Scene Graph Generation by Iterative Message Passing (2017, background): “A series of contextual models based on surrounding pixels and regions have also been developed for perceptual tasks [3, 13, 25, 27].”
• Augmenting Supervised Neural Networks with Unsupervised Objectives for Large-scale Image Classification (2016, background): “Purely supervised learning allowed more flexibility of network architectures, e.g., the inception unit (Szegedy et al., 2015) and the residual structure (He et al., 2016), which were not limited by the modeling assumptions of unsupervised methods.”
• Deep Layer Aggregation (2017, background): “Deep learning diffused across vision by establishing that networks could serve as backbones, which broadcast improvements not once but with every better architecture, through transfer learning [11, 48] and metaalgorithms for object detection [14] and semantic segmentation [35] that take the base arc”
• Zero-Shot Object Detection (2018, methodology): “Early object detection approaches involved getting object proposals for each image and classifying those object proposals using an image classification CNN [16,17,46,55].”
• Evolution of Image Segmentation using Deep Convolutional Neural Network: A Survey (2020, background): “Among different deep learning models, convolutional neural network(CNN) [3, 4, 5] has shown outstanding performance in different high level computer vision task such as image classification [6, 7, 8, 9, 10, 11, 12, 13, 14, 15], object detection [16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29”
• DeepPap: Deep Convolutional Networks for Cervical Cell Classification (2017, methodology): “fer learning [43]–[45] is an effective method to overcome this problem.”
• Text-to-image Synthesis via Symmetrical Distillation Networks (2018, positive): “We can find many successful applications of this paradigm in the various state-of-the-art methods, such as object detection [7], semantic segmentation [21] and fine-grained image classification [13].”
• Diagnosing Error in Temporal Action Detectors (2018, background): “Encouraged by the advances in the image domain through representation learning [14,17,23], large scale datasets have emerged over the last couple of years to challenge existing ideas and enrich our understanding of visual streams [15,16,4,19,21,22,28,37].”
• Learning Rotation-Invariant and Fisher Discriminative Convolutional Neural Networks for Object Detection (2019, methodology): “Our method is implemented based on four successful object detection systems including R-CNN [1], [2], Fast R-CNN [3], Faster R-CNN [4] and R-FCN [5], which are most related with our work and also the baseline methods in our experiments.”
• Fast Convergence of DETR with Spatially Modulated Co-Attention (2021, methodology): “TSP-FCOS and TSP-RCNN [36] combines DETR’s Hungarian matching with FCOS [37] and RCNN [33] detectors, which results in faster convergence and better performance than DETR.”