Indoor Scene Layout Estimation from a Single Image

Lin, Hung Jin; Huang, Sheng Wen; Lai, Shang-Hong; Chiang, Chen-Kuo

doi:10.1109/icpr.2018.8546278

Cited by 22 publications

(18 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Most of the CNN-based approaches for estimating room layout edges employ a encoder-decoder topology with a standard classification network for the encoder and utilize a series of deconvolutional layers for upsampling the feature maps [4], [5], [17], [12]. Ren et al [17] proposed an architecture that employs the VGG-16 network for the encoder followed by fully-connected layers and deconvolutional layers that upsample to one quarter of the input resolution.…”

Section: Related Workmentioning

confidence: 99%

“…The use of fully-connected layers enables their network to have a large receptive field but at the cost of loosing the feature localization ability. Lin et al [4] introduced a similar approach with the stronger ResNet-101 backbone and model the network in a fully-convolutional manner. Most recently, Zhang et al [5] proposed an architecture based on the VGG-16 backbone for simultaneously estimating the layout edges as well as predicting the semantic segmentation of the walls, floor and ceiling.…”

Section: Related Workmentioning

confidence: 99%

“…Employing this gradual thinning approach improves convergence and enables the network to predict precise thin edges, as opposed to training only with a fixed edge width. Lin et al [4] employ a similar training strategy that adaptively changes the edge thickness according to the gradient, however our training strategy resulted in a better performance.…”

Section: B Network Trainingmentioning

confidence: 99%

“…In order to empirically evaluate the performance of our room layout edge extraction network, we performed evaluations on the LSUN benchmark in comparison to state-of-Input Image Lin et al [4] Zhang et al [5] Ours Fig. 6.…”

Section: Evaluation Of Layout Edge Estimationmentioning

confidence: 99%

“…Recently, computationally efficient approaches based on Convolutional Neural Networks (CNNs) have been proposed for extracting structural information from monocular images. This includes methods to extract room layout edges from images [4], [5]. However, these networks occasionally predict discontinuous layout edges, even more in the presence of significant clutter.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Robot Localization in Floor Plans Using a Room Layout Edge Extraction Network

Boniardi

Valada

Mohan

et al. 2019

2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

View full text Add to dashboard Cite

Indoor localization is one of the crucial enablers for deployment of service robots. Although several successful techniques for indoor localization have been proposed, the majority of them relies on maps generated from data gathered with the same sensor modality used for localization. Typically, tedious labor by experts is needed to acquire this data, thus limiting the readiness of the system as well as its ease of installation for inexperienced operators. In this paper, we propose a memory and computationally efficient monocular camera-based localization system that allows a robot to estimate its pose given an architectural floor plan. Our method employs a convolutional neural network to predict room layout edges from a single camera image and estimates the robot pose using a particle filter that matches the extracted edges to the given floor plan. We evaluate our localization system using multiple realworld experiments and demonstrate that it has the robustness and accuracy required for reliable indoor navigation.

show abstract