Deep Structured Instance Graph for Distilling Object Detectors

• URL: http://arxiv.org/abs/2109.12862v1
• Date: Mon, 27 Sep 2021 08:26:00 GMT
• Title: Deep Structured Instance Graph for Distilling Object Detectors
• Authors: Yixin Chen, Pengguang Chen, Shu Liu, Liwei Wang, Jiaya Jia
• Abstract summary: We present a simple knowledge structure to exploit and encode information inside the detection system to facilitate detector knowledge distillation. We achieve new state-of-the-art results on the challenging COCO object detection task with diverse student-teacher pairs on both one- and two-stage detectors.
• Score: 82.16270736573176
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Effectively structuring deep knowledge plays a pivotal role in transfer from teacher to student, especially in semantic vision tasks. In this paper, we present a simple knowledge structure to exploit and encode information inside the detection system to facilitate detector knowledge distillation. Specifically, aiming at solving the feature imbalance problem while further excavating the missing relation inside semantic instances, we design a graph whose nodes correspond to instance proposal-level features and edges represent the relation between nodes. To further refine this graph, we design an adaptive background loss weight to reduce node noise and background samples mining to prune trivial edges. We transfer the entire graph as encoded knowledge representation from teacher to student, capturing local and global information simultaneously. We achieve new state-of-the-art results on the challenging COCO object detection task with diverse student-teacher pairs on both one- and two-stage detectors. We also experiment with instance segmentation to demonstrate robustness of our method. It is notable that distilled Faster R-CNN with ResNet18-FPN and ResNet50-FPN yields 38.68 and 41.82 Box AP respectively on the COCO benchmark, Faster R-CNN with ResNet101-FPN significantly achieves 43.38 AP, which outperforms ResNet152-FPN teacher about 0.7 AP. Code: https://github.com/dvlab-research/Dsig.

Related papers

• Lightweight Monocular Depth Estimation with an Edge Guided Network [34.03711454383413]
  We present a novel lightweight Edge Guided Depth Estimation Network (EGD-Net) In particular, we start out with a lightweight encoder-decoder architecture and embed an edge guidance branch. In order to aggregate the context information and edge attention features, we design a transformer-based feature aggregation module.
  arXiv  Detail & Related papers  (2022-09-29T14:45:47Z)
• Measuring and Sampling: A Metric-guided Subgraph Learning Framework for Graph Neural Network [11.017348743924426]
  We propose a Metric-Guided (MeGuide) subgraph learning framework for Graph neural network (GNN) MeGuide employs two novel metrics: Feature Smoothness and Connection Failure Distance to guide the subgraph sampling and mini-batch based training. We demonstrate the effectiveness and efficiency of MeGuide in training various GNNs on multiple datasets.
  arXiv  Detail & Related papers  (2021-12-30T11:00:00Z)
• Distilling Object Detectors via Decoupled Features [69.62967325617632]
  We present a novel distillation algorithm via decoupled features (DeFeat) for learning a better student detector. Experiments on various detectors with different backbones show that the proposed DeFeat is able to surpass the state-of-the-art distillation methods for object detection.
  arXiv  Detail & Related papers  (2021-03-26T13:58:49Z)
• Node2Seq: Towards Trainable Convolutions in Graph Neural Networks [59.378148590027735]
  We propose a graph network layer, known as Node2Seq, to learn node embeddings with explicitly trainable weights for different neighboring nodes. For a target node, our method sorts its neighboring nodes via attention mechanism and then employs 1D convolutional neural networks (CNNs) to enable explicit weights for information aggregation. In addition, we propose to incorporate non-local information for feature learning in an adaptive manner based on the attention scores.
  arXiv  Detail & Related papers  (2021-01-06T03:05:37Z)
• Pathfinder Discovery Networks for Neural Message Passing [8.633430288397376]
  Pathfinder Discovery Networks (PDNs) are a method for jointly learning a message passing graph over a multiplex network. PDNs inductively learn an aggregated weight for each edge, optimized to produce the best outcome for the downstream learning task.
  arXiv  Detail & Related papers  (2020-10-24T11:28:57Z)
• Information Obfuscation of Graph Neural Networks [96.8421624921384]
  We study the problem of protecting sensitive attributes by information obfuscation when learning with graph structured data. We propose a framework to locally filter out pre-determined sensitive attributes via adversarial training with the total variation and the Wasserstein distance.
  arXiv  Detail & Related papers  (2020-09-28T17:55:04Z)
• Structural Temporal Graph Neural Networks for Anomaly Detection in Dynamic Graphs [54.13919050090926]
  We propose an end-to-end structural temporal Graph Neural Network model for detecting anomalous edges in dynamic graphs. In particular, we first extract the $h$-hop enclosing subgraph centered on the target edge and propose the node labeling function to identify the role of each node in the subgraph. Based on the extracted features, we utilize Gated recurrent units (GRUs) to capture the temporal information for anomaly detection.
  arXiv  Detail & Related papers  (2020-05-15T09:17:08Z)
• Learning to Hash with Graph Neural Networks for Recommender Systems [103.82479899868191]
  Graph representation learning has attracted much attention in supporting high quality candidate search at scale. Despite its effectiveness in learning embedding vectors for objects in the user-item interaction network, the computational costs to infer users' preferences in continuous embedding space are tremendous. We propose a simple yet effective discrete representation learning framework to jointly learn continuous and discrete codes.
  arXiv  Detail & Related papers  (2020-03-04T06:59:56Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.