DAG-Recurrent Neural Networks for Scene Labeling

By Bing Shuai, Zhen Zuo, Bing Wang, G. Wang. 2015

153 citations recorded by the source corpus.

In image labeling, local representations for image units are usually generated from their surrounding image patches, thus long-range contextual information is not effectively encoded. In this paper, we introduce recurrent neural networks (RNNs) to address this issue. Specifically, directed acyclic graph RNNs (DAG-RNNs) are proposed to process DAG-structured images, which enables the network to model long-range semantic dependencies among image units. Our DAG-RNNs are capable of tremendously enhancing the discriminative power of local representations, which significantly benefits the local classification. Meanwhile, we propose a novel class weighting function that attends to rare classes, which phenomenally boosts the recognition accuracy for non-frequent classes. Integrating with convolution and deconvolution layers, our DAG-RNNs achieve new state-of-the-art results on the challenging SiftFlow, CamVid and Barcelona benchmarks.

Subfields: 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 Computer Vision and Pattern Recognition

Citations in the corpus

Recent advances in convolutional neural networks (2015, background): “[189] and [190] use the recurrent neural networks to model the contextual dependencies among image features from CNNs, and dramatically boost the labeling performance.”
Directed Acyclic Graph Network for Conversational Emotion Recognition (2021, background): “, 2015), DAG-RNN(Shuai et al., 2016), D-VAE (Zhang et al.”
Survey on semantic segmentation using deep learning techniques (2019, methodology): “AdapNet [191]: The model is an adapted version of [69], replacing the pooling and convolutional layers of conv4 / conv5 with two dilated convolution layers with dilation factors of 2 and 4 respectively.”
A survey on deep learning techniques for image and video semantic segmentation (2018, methodology): “In particular DAG-RNN is the top scorer with 85.30 IoU.”
Uncertainty-Aware Learning for Zero-Shot Semantic Segmentation (2020, background): “As a fundamental problem in computer vision, this task has attracted a lot of attention from the research community and achieved great success along with the development of deep learning in recent years [7, 8, 14, 15, 19, 24, 38, 45, 50, 51, 57, 59, 60, 63, 66].”
Episodic CAMN: Contextual Attention-Based Memory Networks with Iterative Feedback for Scene Labeling (2017, methodology): “We found that applying class balancing helps similar to [51, 67].”
Topological Recurrent Neural Network for Diffusion Prediction (2017, methodology): “For example, DAG-RNNs are customized for images, which are 2D lattices with ﬁxed orders, instead of a real graph as used in information diffusion.”
Recent Advances in Recurrent Neural Networks (2017, methodology): “A set of images are represented as undirected cyclic graphs (UCGs) in [118].”
A Review on Deep Learning Techniques Applied to Semantic Segmentation (2017, methodology): “Undirected cyclic graphs (UCGs) were also adopted to model image contexts for semantic segmentation [82].”
Leveraging Structural Context Models and Ranking Score Fusion for Human Interaction Prediction (2016, methodology): “Although RNNs and LSTM networks are designed for temporal modeling, they have also been used to process sequences of local features in images to exploit the spatial dependency and extract discriminative information for powerful image representation [16], [33], [34].”
Neural Attribute Machines for Program Generation (2017, background): “While other attempts at learning models that can be used to generate trees include performing convolutions over nodes in binary trees [12] or stacking multiple RNNs in fixed directions [13], a natural paradigm to adopt for presenting a tree to a neural model is a single-traversal, top-down, left-to-”
Distance-Aware DAG Embedding for Proximity Search on Heterogeneous Graphs (2018, unknown): “Becaue the DAGs are directed, we are inspired by the previous work (Sutskever, Vinyals, and Le 2014; Shuai et al. 2016; Zhu, Sobhani, and Guo 2016) to consider recursive and recurrent networks.”