A Multi-Channel and Multi-Spatial Attention Convolutional Neural Network for Prostate Cancer ISUP Grading

Yang, Bochen; Xiao, Zhifeng

doi:10.3390/app11104321

Cited by 9 publications

(7 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Thus, one direction is to simplify the DNN architecture, keeping the model slim and fast. Second, there are other feature extraction techniques to be tested in our system, such as CNN attention modules [59] and feature interaction [60]. Third, a generative data augmentation method can further enrich the vehicle samples in the original dataset.…”

Section: Discussionmentioning

confidence: 99%

RCBi-CenterNet: An Absolute Pose Policy for 3D Object Detection in Autonomous Driving

Chen²,

Wang

et al. 2021

Applied Sciences

Self Cite

View full text Add to dashboard Cite

3D Object detection is a critical mission of the perception system of a self-driving vehicle. Existing bounding box-based methods are hard to train due to the need to remove duplicated detections in the post-processing stage. In this paper, we propose a center point-based deep neural network (DNN) architecture named RCBi-CenterNet that predicts the absolute pose for each detected object in the 3D world space. RCBi-CenterNet is composed of a recursive composite network with a dual-backbone feature extractor and a bi-directional feature pyramid network (BiFPN) for cross-scale feature fusion. In the detection head, we predict a confidence heatmap that is used to determine the position of detected objects. The other pose information, including depth and orientation, is regressed. We conducted extensive experiments on the Peking University/Baidu-Autonomous Driving dataset, which contains more than 60,000 labeled 3D vehicle instances from 5277 real-world images, and each vehicle object is annotated with the absolute pose described by the six degrees of freedom (6DOF). We validated the design choices of various data augmentation methods and the backbone options. Through an ablation study and an overall comparison with the state-of-the-art (SOTA), namely CenterNet, we showed that the proposed RCBi-CenterNet presents performance gains of 2.16%, 2.76%, and 5.24% in Top 1, Top 3, and Top 10 mean average precision (mAP). The model and the result could serve as a credible benchmark for future research in center point-based object detection.

show abstract

Section: Discussionmentioning

confidence: 99%

RCBi-CenterNet: An Absolute Pose Policy for 3D Object Detection in Autonomous Driving

Chen²,

Wang

et al. 2021

Applied Sciences

Self Cite

View full text Add to dashboard Cite

show abstract

“…This work has the following limitations that also point out our future research directions. First, the biattention module applied in the attentive feature extractor only employs a single MLP to learn the attentions, while a multichannel and multispatial attention mod-ule [32] can be adopted for further improvement. Second, the attentive feature extractor only considers one image/feature scale, while a multiscale-based feature pyramid can be utilized to encourage the network to extract features in multiple granularities.…”

Section: Discussionmentioning

confidence: 99%

Learning Attention-Aware Interactive Features for Fine-Grained Vegetable and Fruit Classification

2021

Self Cite

View full text Add to dashboard Cite

Vegetable and fruit recognition can be considered as a fine-grained visual categorization (FGVC) task, which is challenging due to the large intraclass variances and small interclass variances. A mainstream direction to address the challenge is to exploit fine-grained local/global features to enhance the feature extraction and representation in the learning pipeline. However, unlike the human visual system, most of the existing FGVC methods only extract features from individual images during training. In contrast, human beings can learn discriminative features by comparing two different images. Inspired by this intuition, a recent FGVC method, named Attentive Pairwise Interaction Network (API-Net), takes as input an image pair for pairwise feature interaction and demonstrates superior performance in several open FGVC data sets. However, the accuracy of API-Net on VegFru, a domain-specific FGVC data set, is lower than expected, potentially due to the lack of spatialwise attention. Following this direction, we propose an FGVC framework named Attention-aware Interactive Features Network (AIF-Net) that refines the API-Net by integrating an attentive feature extractor into the backbone network. Specifically, we employ a region proposal network (RPN) to generate a collection of informative regions and apply a biattention module to learn global and local attentive feature maps, which are fused and fed into an interactive feature learning subnetwork. The novel neural structure is verified through extensive experiments and shows consistent performance improvement in comparison with the SOTA on the VegFru data set, demonstrating its superiority in fine-grained vegetable and fruit recognition. We also discover that a concatenation fusion operation applied in the feature extractor, along with three top-scoring regions suggested by an RPN, can effectively boost the performance.

show abstract

“…The system achieved an accuracy of 89.4% on the internal test set. In Yang et al (Yang & Xiao 2021), a subset of the tiles are selected based on mean pixel value. A Multi-Channel and Multi-Spatial (MCMS) attention mechanism is proposed to be added to any CNN architecture to improve feature extraction and aid the backbone CNN to focus on more relevant areas of the image.…”

Section: Related Workmentioning

confidence: 99%

Segmentation-Based Fractal Texture Analysis and Deep Convolution Networks for Multigrain Scoring of Prostate Cancer

M.Fathalla¹,

Eltelwany²,

Saad³

2022

INDJCSE

View full text Add to dashboard Cite

Prostate cancer diagnosis and staging is of paramount importance to effective treatment planning and better prognosis. Computer aided diagnosis and staging can contribute to improving and speeding up these stages, especially with the advances of deep learning. The International Society of Urological Pathology (ISUP) grading of stained Whole Slide histopathological Images (WSIs) can be considered the gold standard for grading. However, WSIs suffer from large size and wide background areas which hinder the learning process. Hence, a segmentation-based fractal analysis approach is applied to address this issue and elect relevant patches to be input to the learning algorithm. EfficientNet Convolution Neural Network (CNN) achieves a promising accuracy of 80.7% and Quadratic Weighted Kappa (QWK) of 95.4%. The proposed approach remedies the size problem of WSIs and improves the grading accuracy using light weight learning models.

show abstract

A Multi-Channel and Multi-Spatial Attention Convolutional Neural Network for Prostate Cancer ISUP Grading

Cited by 9 publications

References 28 publications

RCBi-CenterNet: An Absolute Pose Policy for 3D Object Detection in Autonomous Driving

RCBi-CenterNet: An Absolute Pose Policy for 3D Object Detection in Autonomous Driving

Learning Attention-Aware Interactive Features for Fine-Grained Vegetable and Fruit Classification

Segmentation-Based Fractal Texture Analysis and Deep Convolution Networks for Multigrain Scoring of Prostate Cancer

Contact Info

Product

Resources

About