检索结果-内蒙古大学图书馆

Time-domain feature extraction for target-specificity in Photoacoustic Remote Sensing Microscopy

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Pellegrino, Nicholas Ecclestone, Benjamin R. Fieguth, Paul Reza, Parsin Haji PhotoMedicine Labs University of Waterloo 200 University Ave W WaterlooONN2L 3G1 Canada Vision and Image Processing Lab University of Waterloo 200 University Ave W WaterlooONN2L 3G1 Canada IllumiSonics Inc. 22 King Street South Suite 300 WaterlooONN2J 1N8 Canada

Photoacoustic Remote Sensing (PARS) microscopy is an emerging label-free optical absorption imaging modality. PARS operates by capturing nanosecond-scale optical perturbations generated by photoacoustic pressures. These time-domain (TD) modulations are usually projected by amplitude to determine absorption magnitude. However, significant information on the target's material properties is contained within the TD signals. This work proposes a novel clustering method to learn TD features which relate to underlying biomolecule characteristics. This technique identifies features related to constituent biomolecules, enabling single-acquisition virtual tissue labelling. Colorized visualizations of tissue are produced, highlighting specific tissue components. This is demonstrated on freshly resected murine brain tissue, clearly discerning structures including myelinated and unmyelinated neurons (white and gray matter) and nuclear structures. © 2022, CC BY.

关键词： Biomolecules

FactorizeNet: Progressive depth factorization for efficient network architecture exploration under quantization constraints

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Yun, Stone Wong, Alexander University of Waterloo Vision and Image Processing Lab Canada University of Waterloo Waterloo Artificial Intelligence Institute Canada

Depth factorization and quantization have emerged as two of the principal strategies for designing efficient deep convolutional neural network (CNN) architectures tailored for low-power inference on the edge. However, there is still little detailed understanding of how different depth factorization choices affect the final, trained distributions of each layer in a CNN, particularly in the situation of quantized weights and activations. In this study, we introduce a progressive depth factorization strategy for efficient CNN architecture exploration under quantization constraints. By algorithmically increasing the granularity of depth factorization in a progressive manner, the proposed strategy enables a fine-grained, low-level analysis of layer-wise distributions. Thus enabling the gain of in-depth, layer-level insights on efficiency-accuracy tradeoffs under fixed-precision quantization. Such a progressive depth factorization strategy also enables efficient identification of the optimal depth-factorized macroarchitecture design (which we will refer to here as FactorizeNet) based on the desired efficiency-accuracy requirements. © 2020, CC-BY.

关键词： Factorization

30-3: A Moving Camera and Synthetic Calibration Target Solution for Non-Planar Scene Estimation and Projector Calibration

学校读者我要写书评

暂无评论

SID Symposium Digest of Technical Papers 2022年第1期53卷

作者： Katherine Arnold Paul Fieguth Mark Lamm Vision and Image Processing Lab University of Waterloo Waterloo ON Canada Christie Digital Systems Canada Inc. Kitchener ON Canada

The automatic calibration of projector systems improves the possible environments available to this technology which relies on exact scene calibration. This paper explores the challenge of projector calibration and non-planar scene estimation. This formulation assumes no prior information on the moving camera or fixed projector. Limited scene understanding provides the basis for constructing synthetic calibration targets to perform a geometric recovery through bundle adjustment. Synthetic data are generated from ground-truth camera and projector parameters to explore the method's performance over varying properties of the synthetic calibration targets.

关键词： Projector-camera calibration non-planar scene moving camera bundle adjustment auto-calibration self-calibration projection mapping projector calibration

Statistical Shape and Pose Model of the Forearm for Custom Splint Design

学校读者我要写书评

暂无评论

Statistical Shape and Pose Model of the Forearm for Custom S...

IEEE International Symposium on Biomedical Imaging

作者： Femke Danckaers Jeroen Van Houtte Brian G. Booth Frederik Verstreken Jan Sijbers imec - Vision Lab University of Antwerp Antwerp Belgium imec - Image Processing and Interpretation Lab Ghent University Ghent Belgium AZ Monica Hospital Antwerp Belgium

ISBN: (纸本)9781665412469;9781665429474

Custom splint design is becoming more common. However, poor 3D scan quality can negatively impact the design accuracy. This paper describes a method to build a 3D statistical shape and pose model of the forearm from 3dMD scans. The model is used to assist the registration of previously unseen forearms in a wide range of poses. We show that this model-based surface registration results in a good geometric fit, with accurate anatomical correspondences. This method could be used to upgrade low-resolution scans using a high-resolution model.

关键词： Solid modeling Three-dimensional displays Shape Buildings Surface fitting Standards Biomedical imaging

Where does trust break down? A quantitative trust analysis of deep neural networks via trust matrix and conditional trust densities

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Hryniowski, Andrew Wang, Xiao Yu Wong, Alexander Vision and Image Processing Research Group Waterloo Artificial Intelligence Institute University of Waterloo Canada DarwinAI Corp.

The advances and successes in deep learning in recent years have led to considerable efforts and investments into its widespread ubiq uitous adoption for a wide variety of applications, ranging from per sonal assistants and intelligent navigation to search and produc recommendation in e-commerce. With this tremendous rise in deep learning adoption comes questions about the trustworthiness of the deep neural networks that power these applications. Motivated to answer such questions, there has been a very recent interest in trus quantification. In this work, we introduce the concept of trust ma trix, a novel trust quantification strategy that leverages the recentl introduced question-answer trust metric by Wong et al. to provide deeper, more detailed insights into where trust breaks down for a given deep neural network given a set of questions. More specifi cally, a trust matrix defines the expected question-answer trust fo a given actor-oracle answer scenario, allowing one to quickly spo areas of low trust that needs to be addressed to improve the trust worthiness of a deep neural network. The proposed trust matrix i simple to calculate, humanly interpretable, and to the best of the authors’ knowledge is the first to study trust at the actor-oracle an swer level. We further extend the concept of trust densities with the notion of conditional trust densities. We experimentally leverage trust matrices to study several well-known deep neural network ar chitectures for image recognition, and further study the trust densit and conditional trust densities for an interesting actor-oracle answe scenario. The results illustrate that trust matrices, along with condi tional trust densities, can be useful tools in addition to the existing suite of trust quantification metrics for guiding practitioners and reg ulators in creating and certifying deep learning solutions for trusted operation. Copyright © 2020, The Authors. All rights reserved.

关键词： Deep neural networks

Facial Expression Recognition in Videos: An CNN-LSTM based Model for Video Classification

学校读者我要写书评

暂无评论

Facial Expression Recognition in Videos: An CNN-LSTM based M...

International Conference on Electronics, Information and Communications (ICEIC)

作者： Muhammad Abdullah Mobeen Ahmad Dongil Han Vision and Image Processing Lab Sejong University Seoul South Korea

ISBN: (数字)9781728162898

ISBN: (纸本)9781728162904

Facial Expressions are an integral part of human communication. Therefore, correct classification of facial expression in image and video data has been an important quest for researchers and software development industry. In this paper we propose the video classification method using Recurrent Neural Networks (RNN) in addition to Convolution Neural Networks (CNN) to capture temporal as well spatial features of a video sequence. The methodology is tested on The Ryerson Audio-Visual Database of Emotional Speech and Song (RAVDESS). Since no other results were available on this dataset using only visual analysis, the proposed method provides the first benchmark of 61% test accuracy on given dataset.

关键词： convolutional neural nets emotion recognition face recognition feature extraction image capture image classification image sequences recurrent neural nets video signal processing visual databases recurrent neural nets Facial Recognition visual databases image capture Video signal processing emotion recognition image sequences image classification Feature extraction Facial Expression Videotapes expression recognition Personal Communication

AttendNets: tiny deep image recognition neural networks for the edge via visual attention condensers

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Wong, Alexander Famouri, Mahmoud Shafiee, Mohammad Javad Vision and Image Processing Research Group University of Waterloo WaterlooON Canada Waterloo Artificial Intelligence Institute University of Waterloo WaterlooON Canada DarwinAI Corp. WaterlooON Canada

While significant advances in deep learning has resulted in state-of-the-art performance across a large number of complex visual perception tasks, the widespread deployment of deep neural networks for TinyML applications involving on-device, low-power image recognition remains a big challenge given the complexity of deep neural networks. In this study, we introduce AttendNets, low-precision, highly compact deep neural networks tailored for on-device image recognition. More specifically, AttendNets possess deep self-attention architectures based on visual attention condensers, which extends on the recently introduced stand-alone attention condensers to improve spatial-channel selective attention. Furthermore, AttendNets have unique machine-designed macroarchitecture and microarchitecture designs achieved via a machine-driven design exploration strategy. Experimental results on imageNet50 benchmark dataset for the task of on-device image recognition showed that AttendNets have significantly lower architectural and computational complexity when compared to several deep neural networks in research literature designed for efficiency while achieving highest accuracies (with the smallest AttendNet achieving ∼7.2% higher accuracy, while requiring ∼3× fewer multiply-add operations, ∼4.17× fewer parameters, and ∼16.7× lower weight memory requirements than MobileNet-V1). Based on these promising results, AttendNets illustrate the effectiveness of visual attention condensers as building blocks for enabling various on-device visual perception tasks for TinyML applications. Copyright © 2020, The Authors. All rights reserved.

关键词： Deep neural networks

CancerNet-SCa: Tailored deep neural network designs for detection of skin cancer from dermoscopy images

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Ren Hou Lee, James Pavlova, Maya Famouri, Mahmoud Wong, Alexander Vision and Image Processing Research Group University of Waterloo WaterlooON Canada Waterloo Artificial Intelligence Institute University of Waterloo WaterlooON Canada DarwinAI Corp. WaterlooON Canada

Skin cancer continues to be the most frequently diagnosed form of cancer in the U.S., with not only significant effects on health and well-being but also significant economic costs associated with treatment. A crucial step to the treatment and management of skin cancer is effective skin cancer detection due to strong prognosis when treated at an early stage, with one of the key screening approaches being dermoscopy examination. Motivated by the advances of deep learning and inspired by the open source initiatives in the research community, in this study we introduce CancerNet-SCa, a suite of deep neural network designs tailored for the detection of skin cancer from dermoscopy images that is open source and available to the general public as part of the Cancer-Net initiative. To the best of the authors’ knowledge, CancerNet-SCa comprises of the first machine-designed deep neural network architecture designs tailored specifically for skin cancer detection, one of which possessing a self-attention architecture design with attention condensers. Furthermore, we investigate and audit the behaviour of CancerNet-SCa in a responsible and transparent manner via explainability-driven model auditing. While CancerNet-SCa is not a production-ready screening solution, the hope is that the release of CancerNet-SCa in open source, open access form will encourage researchers, clinicians, and citizen data scientists alike to leverage and build upon them. Copyright © 2020, The Authors. All rights reserved.

关键词： Dermatology

Terahertz compressive imaging: understanding and improvement by a better strategy for data selection

学校读者我要写书评

暂无评论

Terahertz compressive imaging: understanding and improvement...

作者： Xing, Chungui Qi, Feng Liu, Zhaoyang Wang, Yelong Guo, Shuxu State Key Laboratory on Integrated Optoelectronics College of Electronic Science and Engineering Jilin University Changchun China Shenyang Institute of Automation Chinese Academy of Sciences Shenyang China Key Laboratory of Opto-Electronic Information Processing Chinese Academy of Sciences Shenyang China Key Lab of Image Understanding and Computer Vision Shenyang China

Compressive sensing (CS) is a novel sampling modality, which indicates the signals can be sampled at a rate much below the Nyquist sampling rate. CS has increasing interest recently due to high demand of rapid, efficient, and in-expensive signal processing applications in the μmWave and mmWave frequencies, such as communication and imaging. There have been a lot of theoretical studies on this topic, but there is a lack of systematic experimental analysis of the implementation method itself. In this paper, we have investigated the influencing factors of terahertz compressive sensing based on experimental results, including illumination and the size of the pixel. Besides, to differentiate from current approaches, which generally make full use of the data, we propose to sort the data first and select a part of them based on amplitude, which might deliver a better image by prompting the mathematical calculations compulsively. We believe that such considerations given above would help to make a better system design and improve the performance of compressive imaging, and these results will also be helpful in the application of terahertz communication. © 2021 John Wiley & Sons, Ltd

关键词： Compressed sensing