2021 Digital Image Computing: Techniques and Applications (DICTA)最新文献_第9页

Quantum Annealing Formulation for Binary Neural Networks 二元神经网络的量子退火公式

2021 Digital Image Computing: Techniques and Applications (DICTA) Pub Date : 2021-07-05 DOI: 10.1109/DICTA52665.2021.9647321

M. Sasdelli, Tat-Jun Chin

{"title":"Quantum Annealing Formulation for Binary Neural Networks","authors":"M. Sasdelli, Tat-Jun Chin","doi":"10.1109/DICTA52665.2021.9647321","DOIUrl":"https://doi.org/10.1109/DICTA52665.2021.9647321","url":null,"abstract":"Quantum annealing is a promising paradigm for building practical quantum computers. Compared to other approaches, quantum annealing technology has been scaled up to a larger number of qubits. On the other hand, deep learning has been profoundly successful in pushing the boundaries of AI. It is thus natural to investigate potentially game changing technologies such as quantum annealers to augment the capabilities of deep learning. In this work, we explore binary neural networks, which are lightweight yet powerful models typically intended for resource constrained devices. Departing from current training regimes for binary networks that smooth/approximate the activation functions to make the network differentiable, we devise a quadratic unconstrained binary optimization formulation for the training problem. While the problem is intractable, i.e., the cost to estimate the binary weights scales exponentially with network size, we show how the problem can be optimized directly on a quantum annealer, thereby opening up to the potential gains of quantum computing. We experimentally validated our formulation via simulation and testing on an actual quantum annealer (D-Wave Advantage), the latter to the extent allowable by the capacity of current technology.","PeriodicalId":424950,"journal":{"name":"2021 Digital Image Computing: Techniques and Applications (DICTA)","volume":"60 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123118463","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 13

EAR-NET: Error Attention Refining Network For Retinal Vessel Segmentation EAR-NET:视网膜血管分割的误差注意细化网络

2021 Digital Image Computing: Techniques and Applications (DICTA) Pub Date : 2021-07-03 DOI: 10.1109/DICTA52665.2021.9647299

Jun Wang, Xiaohan Yu, Yongsheng Gao

{"title":"EAR-NET: Error Attention Refining Network For Retinal Vessel Segmentation","authors":"Jun Wang, Xiaohan Yu, Yongsheng Gao","doi":"10.1109/DICTA52665.2021.9647299","DOIUrl":"https://doi.org/10.1109/DICTA52665.2021.9647299","url":null,"abstract":"The precise detection of blood vessels in retinal images is crucial to the early diagnosis of the retinal vascular diseases, e.g., diabetic, hypertensive and solar retinopathies. Existing works often fail in predicting the abnormal areas, e.g, sudden brighter and darker areas and are inclined to predict a pixel to background due to the significant class imbalance, leading to high accuracy and specificity while low sensitivity. To that end, we propose a novel error attention refining network (ERA-Net) that is capable of learning and predicting the potential false predictions in a two-stage manner for effective retinal vessel segmentation. The proposed ERA-Net in the refine stage drives the model to focus on and refine the segmentation errors produced in the initial training stage. To achieve this, unlike most previous attention approaches that run in an unsupervised manner, we introduce a novel error attention mechanism which considers the differences between the ground truth and the initial segmentation masks as the ground truth to supervise the attention map learning. Experimental results demonstrate that our method achieves state-of-the-art performance on two common retinal blood vessel datasets. Code is available at this link.","PeriodicalId":424950,"journal":{"name":"2021 Digital Image Computing: Techniques and Applications (DICTA)","volume":"40 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128705351","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 7

Multi-Dataset Benchmarks for Masked Identification using Contrastive Representation Learning 使用对比表示学习的屏蔽识别的多数据集基准

2021 Digital Image Computing: Techniques and Applications (DICTA) Pub Date : 2021-06-10 DOI: 10.1109/DICTA52665.2021.9647194

Sachith Seneviratne, Nuran Kasthuriaarachchi, Sanka Rasnayaka

{"title":"Multi-Dataset Benchmarks for Masked Identification using Contrastive Representation Learning","authors":"Sachith Seneviratne, Nuran Kasthuriaarachchi, Sanka Rasnayaka","doi":"10.1109/DICTA52665.2021.9647194","DOIUrl":"https://doi.org/10.1109/DICTA52665.2021.9647194","url":null,"abstract":"The COVID-19 pandemic has drastically changed accepted norms globally. Within the past year, masks have been used as a public health response to limit the spread of the virus. This sudden change has rendered many face recognition based access control, authentication and surveillance systems ineffective. Official documents such as passports, driving license and national identity cards are enrolled with fully uncovered face images. However, in the current global situation, face matching systems should be able to match these reference images with masked face images. As an example, in an airport or security checkpoint it is safer to match the unmasked image of the identifying document to the masked person rather than asking them to remove the mask. We find that current facial recognition techniques are not robust to this form of occlusion. To address this unique requirement presented due to the current circumstance, we propose a set of re-purposed datasets and a benchmark for researchers to use. We also propose a contrastive visual representation learning based pre-training workflow which is specialized to masked vs unmasked face matching. We ensure that our method learns robust features to differentiate people across varying data collection scenarios. We achieve this by training over many different datasets and validating our result by testing on various holdout datasets, including a real world dataset collected specifically for evaluation. The specialized weights trained by our method outperform standard face recognition features for masked to unmasked face matching. We believe the provided synthetic mask generating code, our novel training approach and the trained weights from the masked face models will help in adopting existing face recognition systems to operate in the current global environment. We open-source all contributions for broader use by the research community.","PeriodicalId":424950,"journal":{"name":"2021 Digital Image Computing: Techniques and Applications (DICTA)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-06-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129609521","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 5

Elimination of Central Artefacts of L-SPECT with Modular Partial Ring Detectors by Shifting Center of Scanning 模块化局部环检测器扫描中心移位去除L-SPECT中心伪影

2021 Digital Image Computing: Techniques and Applications (DICTA) Pub Date : 2020-08-18 DOI: 10.1109/DICTA52665.2021.9647107

Manu Francis, M. Tahtali, M. Pickering

{"title":"Elimination of Central Artefacts of L-SPECT with Modular Partial Ring Detectors by Shifting Center of Scanning","authors":"Manu Francis, M. Tahtali, M. Pickering","doi":"10.1109/DICTA52665.2021.9647107","DOIUrl":"https://doi.org/10.1109/DICTA52665.2021.9647107","url":null,"abstract":"The Lightfield Single Photon Emission Computed Tomography (L-SPECT) system is developed to overcome some of the drawbacks in conventional SPECT by applying the idea of plenoptic imaging. This system displayed improved performance in terms of reduced information loss and scanning time compared to the SPECT system, which has a conventional collimator. The SPECT system is transformed into an L-SPECT system by replacing the conventional collimators with micro-range multi-pinhole arrays. The L-SPECT system's resolution is enhanced by reshaping the detector head into ring-type by tiling small detector modules. The L-SPECT system with modular partial ring detectors (MPRD L-SPECT) exhibits cylindrical artefacts during volumetric reconstruction. Hence, here the work is focused to remove the cylindrical artefacts in the reconstruction of the examined objects by changing the scanning orbit. The enhancement is done such that the center of scanning of the L-SPECT system with MPRD L-SPECT is shifted at different values. The reconstruction quality of MPRD L-SPECT with and without center shifting is evaluated in terms of the Full Width at Half Maximum (FWHM) and Modulation Transfer Function (MTF). Moreover the visual comparison is also examined. The results indicate that center shifting of MPRD L-SPECT overcomes the problem of the central artefact with improved FWHM. By increasing MPRD's scanning centre shift gap, the spatial resolution can be further improved.","PeriodicalId":424950,"journal":{"name":"2021 Digital Image Computing: Techniques and Applications (DICTA)","volume":"189 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-08-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131892423","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Automatic Pruning for Quantized Neural Networks 量化神经网络的自动修剪

2021 Digital Image Computing: Techniques and Applications (DICTA) Pub Date : 2020-02-03 DOI: 10.1109/DICTA52665.2021.9647074

Luis Guerra, Bohan Zhuang, I. Reid, T. Drummond

引用次数: 17

Semi-supervised Learning via Conditional Rotation Angle Estimation 基于条件旋转角度估计的半监督学习

2021 Digital Image Computing: Techniques and Applications (DICTA) Pub Date : 2020-01-09 DOI: 10.1109/DICTA52665.2021.9647327

Hai-Ming Xu, Lingqiao Liu, Dong Gong

{"title":"Semi-supervised Learning via Conditional Rotation Angle Estimation","authors":"Hai-Ming Xu, Lingqiao Liu, Dong Gong","doi":"10.1109/DICTA52665.2021.9647327","DOIUrl":"https://doi.org/10.1109/DICTA52665.2021.9647327","url":null,"abstract":"Self-supervised learning (SlfSL), aiming at learning feature representations through ingeniously designed pretext tasks without human annotation, has achieved compelling progress in the past few years. Very recently, SlfSL has also been identified as a promising solution for semi-supervised learning (SemSL) since it offers a new paradigm to utilize unlabeled data. This work further explores this direction by proposing to couple SlfSL with SemSL. Our insight is that the prediction target in SemSL can be modeled as the latent factor in the predictor for the SlfSL target. Marginalizing over the latent factor naturally derives a new formulation which marries the prediction targets of these two learning processes. By implementing this idea through a simple-but-effective SlfSL approach - rotation angle prediction, we create a new SemSL approach called Conditional Rotation Angle EStimation (CRAES). Specifically, CRAES is featured by adopting a module which predicts the image rotation angle conditioned on the candidate image class. Through experimental evaluation, we show that CRAES achieves superior performance over the other existing ways of combining SlfSL and SemSL. To further boost CRAES, we propose two extensions to strengthen the coupling between SemSL target and SlfSL target in basic CRAES. We show that this leads to an improved CRAES method which can achieve the state-of-the-art SemSL performance.","PeriodicalId":424950,"journal":{"name":"2021 Digital Image Computing: Techniques and Applications (DICTA)","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-01-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114584901","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 2

Robust Re-identification of Manta Rays from Natural Markings by Learning Pose Invariant Embeddings 基于姿态不变嵌入的自然标记对蝠鲼的鲁棒再识别

2021 Digital Image Computing: Techniques and Applications (DICTA) Pub Date : 2019-02-28 DOI: 10.1109/DICTA52665.2021.9647359

Olga Moskvyak, F. Maire, A. Armstrong, Feras Dayoub, Mahsa Baktash

{"title":"Robust Re-identification of Manta Rays from Natural Markings by Learning Pose Invariant Embeddings","authors":"Olga Moskvyak, F. Maire, A. Armstrong, Feras Dayoub, Mahsa Baktash","doi":"10.1109/DICTA52665.2021.9647359","DOIUrl":"https://doi.org/10.1109/DICTA52665.2021.9647359","url":null,"abstract":"Visual re-identification of individual animals that bear unique natural body markings is an essential task in wildlife conservation. The photo databases of animal markings grow with each new observation and identifying an individual means matching against thousands of images. We focus on the re-identification of manta rays because the existing process is time-consuming and only semi-automatic. The current solution Manta Matcher requires images of high quality with the pattern of interest in a near frontal view limiting the use of photos sourced from citizen scientists. This paper presents a novel application of a deep convolutional neural network (CNN) for visual re-identification based on natural markings. Our contribution is an experimental demonstration of the superiority of CNNs in learning embeddings for patterns under viewpoint changes on a novel and challenging dataset. We show that our system can handle more variations in viewing angle, occlusions and illumination compared to the current solution. Our system achieves top-10 accuracy of 98% with only 2 matching examples in the database which makes it of practical value and ready for adoption by marine biologists. We also evaluate our system on a dataset of humpback whale flukes to demonstrate that the approach is generic and not species-specific.","PeriodicalId":424950,"journal":{"name":"2021 Digital Image Computing: Techniques and Applications (DICTA)","volume":"47 ","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-02-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"120931296","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 28