Pit-Pattern Classification of Colorectal Cancer Polyps Using a Hyper Sensitive Vision-Based Tactile Sensor and Dilated Residual Networks

• URL: http://arxiv.org/abs/2211.06814v1
• Date: Sun, 13 Nov 2022 04:42:10 GMT
• Title: Pit-Pattern Classification of Colorectal Cancer Polyps Using a Hyper Sensitive Vision-Based Tactile Sensor and Dilated Residual Networks
• Authors: Nethra Venkatayogi, Qin Hu, Ozdemir Can Kara, Tarunraj G. Mohanraj, S. Farokh Atashzar, Farshid Alambeigi
• Abstract summary: We propose utilizing a hyper-sensitive vision-based tactile sensor called HySenSe and a complementary and novel machine learning architecture. The proposed architecture was compared with the state-of-the-art ML models (e.g., AlexNet and DenseNet) and proved to be superior in terms of performance and complexity.
• Score: 4.056583163276972
• License: http://creativecommons.org/licenses/by-nc-nd/4.0/
• Abstract: In this study, with the goal of reducing the early detection miss rate of colorectal cancer (CRC) polyps, we propose utilizing a novel hyper-sensitive vision-based tactile sensor called HySenSe and a complementary and novel machine learning (ML) architecture that explores the potentials of utilizing dilated convolutions, the beneficial features of the ResNet architecture, and the transfer learning concept applied on a small dataset with the scale of hundreds of images. The proposed tactile sensor provides high-resolution 3D textural images of CRC polyps that will be used for their accurate classification via the proposed dilated residual network. To collect realistic surface patterns of CRC polyps for training the ML models and evaluating their performance, we first designed and additively manufactured 160 unique realistic polyp phantoms consisting of 4 different hardness. Next, the proposed architecture was compared with the state-of-the-art ML models (e.g., AlexNet and DenseNet) and proved to be superior in terms of performance and complexity.

Related papers

• GTR: Improving Large 3D Reconstruction Models through Geometry and Texture Refinement [51.97726804507328]
  We propose a novel approach for 3D mesh reconstruction from multi-view images. Our method takes inspiration from large reconstruction models that use a transformer-based triplane generator and a Neural Radiance Field (NeRF) model trained on multi-view images.
  arXiv  Detail & Related papers  (2024-06-09T05:19:24Z)
• ECC-PolypDet: Enhanced CenterNet with Contrastive Learning for Automatic Polyp Detection [88.4359020192429]
  Existing methods either involve computationally expensive context aggregation or lack prior modeling of polyps, resulting in poor performance in challenging cases. In this paper, we propose the Enhanced CenterNet with Contrastive Learning (ECC-PolypDet), a two-stage training & end-to-end inference framework. Box-assisted Contrastive Learning (BCL) during training to minimize the intra-class difference and maximize the inter-class difference between foreground polyps and backgrounds, enabling our model to capture concealed polyps. In the fine-tuning stage, we introduce the IoU-guided Sample Re-weighting
  arXiv  Detail & Related papers  (2024-01-10T07:03:41Z)
• A novel method to enhance pneumonia detection via a model-level ensembling of CNN and vision transformer [0.7499722271664147]
  Pneumonia remains a leading cause of morbidity and mortality worldwide. Deep learning has shown immense potential for pneumonia detection from Chest X-ray (CXR) imaging. We developed a novel model fusing Convolution Neural networks (CNN) and Vision Transformer networks via model-level ensembling.
  arXiv  Detail & Related papers  (2024-01-04T16:58:31Z)
• Visual Prompting Upgrades Neural Network Sparsification: A Data-Model Perspective [67.25782152459851]
  We introduce a novel data-model co-design perspective: to promote superior weight sparsity. Specifically, customized Visual Prompts are mounted to upgrade neural Network sparsification in our proposed VPNs framework.
  arXiv  Detail & Related papers  (2023-12-03T13:50:24Z)
• ResFields: Residual Neural Fields for Spatiotemporal Signals [61.44420761752655]
  ResFields is a novel class of networks specifically designed to effectively represent complex temporal signals. We conduct comprehensive analysis of the properties of ResFields and propose a matrix factorization technique to reduce the number of trainable parameters. We demonstrate the practical utility of ResFields by showcasing its effectiveness in capturing dynamic 3D scenes from sparse RGBD cameras.
  arXiv  Detail & Related papers  (2023-09-06T16:59:36Z)
• Towards Reliable Colorectal Cancer Polyps Classification via Vision Based Tactile Sensing and Confidence-Calibrated Neural Networks [1.4518956926610687]
  We develop a residual neural network to address its over-confident outputs for CRC polyps classification. We introduce noise and blur to the obtained textural images of the vision-based tactile sensing (VS-TS) system and test the model's reliability for non-ideal inputs.
  arXiv  Detail & Related papers  (2023-04-25T23:18:13Z)
• Improving the Timing Resolution of Positron Emission Tomography Detectors Using Boosted Learning -- A Residual Physics Approach [0.4999814847776097]
  This work presents a novel and versatile approach to detector optimization using machine learning (ML) and residual physics. We apply the concept to positron emission tomography (PET), intending to improve the coincidence time resolution (CTR) We were able to improve the CTR significantly (more than 20%) for clinically relevant detectors of 19 mm height, reaching CTRs of 185 ps (450-550 keV)
  arXiv  Detail & Related papers  (2023-02-03T12:10:24Z)
• Classification of Colorectal Cancer Polyps via Transfer Learning and Vision-Based Tactile Sensing [0.2446672595462589]
  We explore the potentials of utilizing transfer learning and machine learning (ML) classifiers to precisely and classify the type of colorectal cancer (CRC) polyps. Instead of using the common colonoscopic images, we applied three different ML algorithms on the 3D textural image outputs of a unique vision-based surface tactile sensor (VS-TS) The performance of the used three ML algorithms in classifying the type of fabricated polyps was quantitatively evaluated using various statistical metrics.
  arXiv  Detail & Related papers  (2022-11-08T21:47:36Z)
• An Adversarial Active Sampling-based Data Augmentation Framework for Manufacturable Chip Design [55.62660894625669]
  Lithography modeling is a crucial problem in chip design to ensure a chip design mask is manufacturable. Recent developments in machine learning have provided alternative solutions in replacing the time-consuming lithography simulations with deep neural networks. We propose a litho-aware data augmentation framework to resolve the dilemma of limited data and improve the machine learning model performance.
  arXiv  Detail & Related papers  (2022-10-27T20:53:39Z)
• Learning to Synthesize Volumetric Meshes from Vision-based Tactile Imprints [26.118805500471066]
  Vision-based tactile sensors typically utilize a deformable elastomer and a camera mounted above to provide high-resolution image observations of contacts. This paper focuses on learning to synthesize the mesh of the elastomer based on the image imprints acquired from vision-based tactile sensors. A graph neural network (GNN) is introduced to learn the image-to-mesh mappings with supervised learning.
  arXiv  Detail & Related papers  (2022-03-29T00:24:10Z)
• Hierarchical Amortized Training for Memory-efficient High Resolution 3D GAN [52.851990439671475]
  We propose a novel end-to-end GAN architecture that can generate high-resolution 3D images. We achieve this goal by using different configurations between training and inference. Experiments on 3D thorax CT and brain MRI demonstrate that our approach outperforms state of the art in image generation.
  arXiv  Detail & Related papers  (2020-08-05T02:33:04Z)

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.