Light-weight visual place recognition using convolutional neural network for mobile robots

Park, Chanjong; Jang, Junik; Zhang, Lei; Jung, Ji‐Hye

doi:10.1109/icce.2018.8326274

Cited by 7 publications

(5 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It is clear that the CNN-based methods outperform the previous works. Based on CNN networks, researchers proposed many methods for scene recognition, such as graph-based CNN [18], light-weight CNN [19] and VLAD-based CNN [20]. The graph-based CNN is constructed by combining the features extracted from CNN and the temporal information of the images in a sequence, and the graph just includes nodes and edges, which greatly reduces computational consumption.…”

Section: Cnn-based Methodsmentioning

confidence: 99%

An End-to-End Trainable Multi-Column CNN for Scene Recognition in Extremely Changing Environment

Zhou

Shen

2020

Sensors

View full text Add to dashboard Cite

Scene recognition is an essential part in the vision-based robot navigation domain. The successful application of deep learning technology has triggered more extensive preliminary studies on scene recognition, which all use extracted features from networks that are trained for recognition tasks. In the paper, we interpret scene recognition as a region-based image retrieval problem and present a novel approach for scene recognition with an end-to-end trainable Multi-column convolutional neural network (MCNN) architecture. The proposed MCNN utilizes filters with receptive fields of different sizes to have Multi-level and Multi-layer image perception, and consists of three components: front-end, middle-end and back-end. The first seven layers VGG16 are taken as front-end for two-dimensional feature extraction, Inception-A is taken as the middle-end for deeper learning feature representation, and Large-Margin Softmax Loss (L-Softmax) is taken as the back-end for enhancing intra-class compactness and inter-class-separability. Extensive experiments have been conducted to evaluate the performance according to compare our proposed network to existing state-of-the-art methods. Experimental results on three popular datasets demonstrate the robustness and accuracy of our approach. To the best of our knowledge, the presented approach has not been applied for the scene recognition in literature.

show abstract

Section: Cnn-based Methodsmentioning

confidence: 99%

An End-to-End Trainable Multi-Column CNN for Scene Recognition in Extremely Changing Environment

Zhou

Shen

2020

Sensors

View full text Add to dashboard Cite

show abstract

“…Sun et al 18 proposed a novel point-cloud-based place recognition method using a pretrained CNN. Park et al 19 proposed a lightweight CNN place recognition method for embedded systems, and the experimental results show that the method is significantly better than traditional methods in accuracy and calculation time. By combining the deep learning features with spatiotemporal filtering, Chen et al 20 completed the place recognition task based on CNN models.…”

Section: Related Workmentioning

confidence: 99%

Robust place recognition based on salient landmarks screening and convolutional neural network features

Niu

Qian

2020

International Journal of Advanced Robotic Systems

View full text Add to dashboard Cite

In this work, we propose a robust place recognition measurement in natural environments based on salient landmark screening and convolutional neural network (CNN) features. First, the salient objects in the image are segmented as candidate landmarks. Then, a category screening network is designed to remove specific object types that are not suitable for environmental modeling. Finally, a three-layer CNN is used to get highly representative features of the salient landmarks. In the similarity measurement, a Siamese network is chosen to calculate the similarity between images. Experiments were conducted on three challenging benchmark place recognition datasets and superior performance was achieved compared to other state-of-the-art methods, including FABMAP, SeqSLAM, SeqCNNSLAM, and PlaceCNN. Our method obtains the best results on the precision–recall curves, and the average precision reaches 78.43%, which is the best of the comparison methods. This demonstrates that the CNN features on the screened salient landmarks can be against a strong viewpoint and condition variations.

show abstract

“…In their work, depending on the appearance of every position, suggested the probabilistic way of place recognition. (Park et al, 2018) Presented a light-weight visual place recognition which is depending on CNN. The presented architecture is particularly designed for mobile robots which supplied with embedded systems.…”

Section: Related Workmentioning

confidence: 99%

Indoor Place Recognition and Localization Using Histogram of Oriented Gradient with Deep Learning

2020

ZJPAS

View full text Add to dashboard Cite

Indoor place recognition is a crucial and challenging field of computer science. It is widely used in robotics and computer vision for various applications. The challenges in indoor place recognition comes from the fact that recognizing localized places like office, corridor, and others may fall under various environmental effects of weather, illumination and others. In this paper, an indoor place recognition and localization system is proposed. The system utilizes the great recognition capabilities of Convolutional Neural Network (CNN) and AlexNet with the use of feature image for training. The feature images are constructed using Histogram of Gradient (HOG). The main contribution of this work is the use of 2D feature constructed image from HOG instead of the scene image used with CNN. The proposed system was compared to other previous systems, in which, it achieved better recognition accuracy when tested on COLD and IDOL standard indoor image datasets.

show abstract

Light-weight visual place recognition using convolutional neural network for mobile robots

Cited by 7 publications

References 16 publications

An End-to-End Trainable Multi-Column CNN for Scene Recognition in Extremely Changing Environment

An End-to-End Trainable Multi-Column CNN for Scene Recognition in Extremely Changing Environment

Robust place recognition based on salient landmarks screening and convolutional neural network features

Indoor Place Recognition and Localization Using Histogram of Oriented Gradient with Deep Learning

Contact Info

Product

Resources

About