Filters








30,464 Hits in 4.7 sec

Deep Bottleneck Feature for Image Classification

Yan Song, Ian McLoughLin, Lirong Dai
2015 Proceedings of the 5th ACM on International Conference on Multimedia Retrieval - ICMR '15  
In this paper, we propose a bag of Deep Bottleneck Features (DBF) for image classification, effectively combining the strengths of a CNN within a BoF framework.  ...  Effective image representation plays an important role for image classification and retrieval. Bag-of-Features (BoF) is well known as an effective and robust visual representation.  ...  Section 2 will introduce a BoDBF framework for image classification.  ... 
doi:10.1145/2671188.2749314 dblp:conf/mir/SongMD15 fatcat:s3nzs7srbfgbblizxge3xeuz5e

Deep complementary bottleneck features for visual speech recognition

Stavros Petridis, Maja Pantic
2016 2016 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)  
Deep bottleneck features (DBNFs) have been used successfully in the past for acoustic speech recognition from audio.  ...  We first train a deep autoencoder with a bottleneck layer in order to reduce the dimensionality of the image.  ...  INTRODUCTION Deep bottleneck features (DBNFs) have been used successfully for acoustic speech recognition.  ... 
doi:10.1109/icassp.2016.7472088 dblp:conf/icassp/PetridisP16 fatcat:ut5tzskabrflzk6mg7t2q6dqpe

End-To-End Visual Speech Recognition With LSTMs [article]

Stavros Petridis, Zuwei Li, Maja Pantic
2017 arXiv   pre-print
Recently, several deep learning approaches have been presented which automatically extract features from the mouth images and aim to replace the feature extraction stage.  ...  The model consists of two streams which extract features directly from the mouth and difference images, respectively.  ...  In our previous work [11] , we extracted bottleneck features directly from the raw mouth ROI using a deep feedforward network and then trained an LSTM for classification. Noda et al.  ... 
arXiv:1701.05847v1 fatcat:olqbhpwjvjhf7mdvoc5j5cs674

End-to-end visual speech recognition with LSTMS

Stavros Petridis, Zuwei Li, Maja Pantic
2017 2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)  
Recently, several deep learning approaches have been presented which automatically extract features from the mouth images and aim to replace the feature extraction stage.  ...  The model consists of two streams which extract features directly from the mouth and difference images, respectively.  ...  In our previous work [11] , we extracted bottleneck features directly from the raw mouth ROI using a deep feedforward network and then trained an LSTM for classification. Noda et al.  ... 
doi:10.1109/icassp.2017.7952625 dblp:conf/icassp/PetridisLP17 fatcat:vg2mewdsibgflktoeibzu6fvrq

Accelerating Deep Neural Networks with Spatial Bottleneck Modules [article]

Junran Peng, Lingxi Xie, Zhaoxiang Zhang, Tieniu Tan, Jingdong Wang
2018 arXiv   pre-print
This paper presents an efficient module named spatial bottleneck for accelerating the convolutional layers in deep neural networks.  ...  We empirically verify the effectiveness of spatial bottleneck by applying it to the deep residual networks.  ...  We empirically verify that spatial bottleneck achieves comparable accuracies in CIFAR image classification and even higher accuracies in ImageNet image classification.  ... 
arXiv:1809.02601v1 fatcat:rocesl6lorcylbzvhxfjbxawf4

Automated Diagnosis of Childhood Pneumonia in Chest Radiographs Using Modified Densely Residual Bottleneck-Layer Features

Sinan Alkassar, Mohammed A. M. Abdullah, Bilal A. Jebur, Ghassan H. Abdul-Majeed, Bo Wei, Wai Lok Woo
2021 Applied Sciences  
few light-weighted densely connected bottleneck residual block features to extract rich spatial information.  ...  Next, an adaptive weight setup is proposed utilizing Adaboost ensemble learning which adaptively sets weight for each classifier depending on the scores generated to achieve the highest true positive rates  ...  The confusion matrix of the proposed model in a normal vs. bacterial vs. viral CXR images classification fashion using (a) deep features and (b) proposed features.  ... 
doi:10.3390/app112311461 fatcat:ecdqwfzrybdgfamge7egv4acqu

Deep Learning based Approach for Bone Diagnosis Classification in Ultrasonic Computed Tomographic Images

Marwa Fradi, Mouna Afif, Mohsen Machhout
2020 International Journal of Advanced Computer Science and Applications  
At second step, an evolutionary neural network is proposed with the AmeobaNet model for USCT image classification.  ...  Results achieve 100% for train accuracy and 96%, 91.7% and 87.5% using Amoebanet, Inception-V3 and MobileNet respectively for the test accuracy.  ...  In [11] , authors have employed a deep model and statistic feature fusion for feature extraction with a multilayer perceptron for medical image classification giving high classification results.  ... 
doi:10.14569/ijacsa.2020.0111210 fatcat:guz65s3ii5b3zib6o7yryvcnqa

sj-pdf-1-jcb-10.1177_0271678X221129230 - Supplemental material for Interpretable deep learning for the prognosis of long-term functional outcome post-stroke using acute diffusion weighted imaging

Eric Moulton, Romain Valabregue, Michel Piotin, Gaultier Marnat, Suzana Saleme, Bertrand Lapergue, Stephane Lehericy, Frederic Clarencon, Charlotte Rosso
2022 Figshare  
Supplemental material, sj-pdf-1-jcb-10.1177_0271678X221129230 for Interpretable deep learning for the prognosis of long-term functional outcome post-stroke using acute diffusion weighted imaging by Eric  ...  In doing so, the network is able to highlight and attend more to regions of the image containing importantsemantic features crucial to proper classification.  ...  ] [0.72-0.83] bottleneck Table 4 : Classification Performance Per Site / Fold for the Ensemble Model. 4 Test Site N Accuracy AUC Sensitivity Specificity PPV NPV Site 1 113 0.79 0.85 0.56 0.93 0.83 0.77  ... 
doi:10.25384/sage.21230342.v1 fatcat:e5ueaz7igndjnlyf3c7fqw7anu

A Deep Bottleneck U-Net Combined With Saliency Map For Classifying Diabetic Retinopathy In Fundus Images

Vo Thi Hong Tuyet, Nguyen Thanh Binh, Dang Thanh Tin
2022 International Journal of Online and Biomedical Engineering (iJOE)  
This paper proposes a method for classification of diabetic retinopathy using saliency and shape detection of objects based on a deep Bottleneck U-Net (DbU-Net) and support vector machines in retinal blood  ...  Classification of diabetic retinopathy in fundus images is very challenging because the blood vessels in the retinal images are too small.  ...  We acknowledge the support of time and facilities from Ho Chi Minh City University of Technology (HCMUT), VNU-HCM for this study.  ... 
doi:10.3991/ijoe.v18i02.27605 fatcat:k3gq6tx26vhdtbj3quzvi5a46m

Tracking And Identifying Floating Marine Debris

Kylili Kyriaki, Artusi Alessandro, Kyriakides Ioannis, Hadjistassou Constantinos
2018 Zenodo  
CNN using Bottleneck to create classifiers able to classify images of marine debris in respective categories.  ...  The main goal of this research is to devise an automatic method for detecting and classifying marine debris. To this end, we are using two machine learning methods: A. Bag of Features B.  ...  Testing the two methods on the same image set we report that the Deep Learning approach reaches a classification accuracy of 77.5% compared to 62.5% for the Bag of Features.  ... 
doi:10.5281/zenodo.1227715 fatcat:nvlglev2uzheddfzwpehzs3omy

End-to-End Visual Speech Recognition for Small-Scale Datasets [article]

Stavros Petridis, Yujiang Wang, Pingchuan Ma, Zuwei Li, Maja Pantic
2019 arXiv   pre-print
Several deep learning approaches have been recently presented aiming to replace the feature extraction stage by automatically extracting features from mouth images.  ...  However, research on joint learning of features and classification remains limited.  ...  In the second generation of deep models, deep bottleneck architectures were used which extract bottleneck features directly from the pixels.  ... 
arXiv:1904.01954v4 fatcat:tk3rbdehk5h7rnhxfy5hhccao4

Residual Dense Network Based on Channel-Spatial Attention for the Scene Classification of a High-Resolution Remote Sensing Image

Xiaolei Zhao, Jing Zhang, Jimiao Tian, Li Zhuo, Jie Zhang
2020 Remote Sensing  
Therefore, we design an RDN based on channel-spatial attention for scene classification of a high-resolution remote sensing image.  ...  The scene classification of a remote sensing image has been widely used in various fields as an important task of understanding the content of a remote sensing image.  ...  Acknowledgments: The authors would like to thank the anonymous reviewers and associate editor for their valuable comments and suggestions to improve the quality of the paper.  ... 
doi:10.3390/rs12111887 fatcat:zcebi6ahxjavbcpowtbsetnu5a

Dog Breed Identification with Fine tuning of Pre-trained models

2019 International journal of recent technology and engineering  
and obtain bottleneck features from these pre-trained models.  ...  Convolutional Neural Networks requires a large amount of images as training data and basic time for training the data and to achieve higher accuracy on the classification.  ...  CONCLUSION This paper describes the fine tuning of bottleneck features extracted through transfer learning by using multiple pretrained deep CNN"s and produced better results compared to traditional classification  ... 
doi:10.35940/ijrte.b1464.0982s1119 fatcat:zhcmwr4uz5gmfha65c4wcekgfi

Image semantic segmentation method based on improved ERFNet model

Dexue Ye, Rubing Han
2021 The Journal of Engineering  
Then, global pooling is used to fuse the feature channels after pyramid pooling to preserve more important feature information.  ...  , an image semantic segmentation method based on improved ERFNet model is proposed.  ...  Literature [23] proposed a semantic segmentation method for traffic scenes based on RGB-D images and deep learning.  ... 
doi:10.1049/tje2.12104 fatcat:eabzc4xazvaurdarbyxy3qyzom

Design and Implementation of Deep Learning Based Contactless Authentication System Using Hand Gestures

Aveen Dayal, Naveen Paluru, Linga Reddy Cenkeramaddi, Soumya J., Phaneendra K. Yalavarthy
2021 Electronics  
The proposed deep learning model is based on the bottleneck module which is inspired by the deep residual networks.  ...  The model achieves classification accuracy of 99.1% on the publicly available sign language digits dataset.  ...  Acknowledgments: We would like to thank the Norwegian Research Council for the support through the INCAPS project: 287918 of the INTPART program.  ... 
doi:10.3390/electronics10020182 fatcat:ntkc5wsoajefhe34gvmzupwnye
« Previous Showing results 1 — 15 out of 30,464 results