scholarly journals Hierarchical Domain-Adapted Feature Learning for Video Saliency Prediction

2021 ◽  
Author(s):  
G. Bellitto ◽  
F. Proietto Salanitri ◽  
S. Palazzo ◽  
F. Rundo ◽  
D. Giordano ◽  
...  
Keyword(s):  
Multiple Scales ◽  
Domain Adaptation ◽  
State Of The Art ◽  
Feature Learning ◽  
Hierarchical Learning ◽  
Domain Specific ◽  
Proposed Model ◽  
Saliency Prediction ◽  
Video Saliency ◽  
Abstraction Levels

AbstractIn this work, we propose a 3D fully convolutional architecture for video saliency prediction that employs hierarchical supervision on intermediate maps (referred to as conspicuity maps) generated using features extracted at different abstraction levels. We provide the base hierarchical learning mechanism with two techniques for domain adaptation and domain-specific learning. For the former, we encourage the model to unsupervisedly learn hierarchical general features using gradient reversal at multiple scales, to enhance generalization capabilities on datasets for which no annotations are provided during training. As for domain specialization, we employ domain-specific operations (namely, priors, smoothing and batch normalization) by specializing the learned features on individual datasets in order to maximize performance. The results of our experiments show that the proposed model yields state-of-the-art accuracy on supervised saliency prediction. When the base hierarchical model is empowered with domain-specific modules, performance improves, outperforming state-of-the-art models on three out of five metrics on the DHF1K benchmark and reaching the second-best results on the other two. When, instead, we test it in an unsupervised domain adaptation setting, by enabling hierarchical gradient reversal layers, we obtain performance comparable to supervised state-of-the-art. Source code, trained models and example outputs are publicly available at https://github.com/perceivelab/hd2s.

scholarly journals Domain Conditioned Adaptation Network

2020 ◽  
Vol 34 (07) ◽  
pp. 11386-11393 ◽  
Author(s):  
Shuang Li ◽  
Chi Liu ◽  
Qiuxia Lin ◽  
Binhui Xie ◽  
Zhengming Ding ◽  
...  
Keyword(s):  
Domain Adaptation ◽  
Feature Learning ◽  
Feature Representations ◽  
Domain Specific ◽  
Cross Domain ◽  
Learning Tasks ◽  
Domain Learning ◽  
High Level ◽  
High Level Feature ◽  
Target Data

Tremendous research efforts have been made to thrive deep domain adaptation (DA) by seeking domain-invariant features. Most existing deep DA models only focus on aligning feature representations of task-specific layers across domains while integrating a totally shared convolutional architecture for source and target. However, we argue that such strongly-shared convolutional layers might be harmful for domain-specific feature learning when source and target data distribution differs to a large extent. In this paper, we relax a shared-convnets assumption made by previous DA methods and propose a Domain Conditioned Adaptation Network (DCAN), which aims to excite distinct convolutional channels with a domain conditioned channel attention mechanism. As a result, the critical low-level domain-dependent knowledge could be explored appropriately. As far as we know, this is the first work to explore the domain-wise convolutional channel activation for deep DA networks. Moreover, to effectively align high-level feature distributions across two domains, we further deploy domain conditioned feature correction blocks after task-specific layers, which will explicitly correct the domain discrepancy. Extensive experiments on three cross-domain benchmarks demonstrate the proposed approach outperforms existing methods by a large margin, especially on very tough cross-domain learning tasks.

scholarly journals Semantic Locality-Aware Deformable Network for Clothing Segmentation

2018 ◽  
Author(s):  
Wei Ji ◽  
Xi Li ◽  
Yueting Zhuang ◽  
Omar El Farouk Bourahla ◽  
Yixin Ji ◽  
...  
Keyword(s):  
Learning Process ◽  
State Of The Art ◽  
Semantic Segmentation ◽  
Small Sample ◽  
Sample Problem ◽  
Fine Grained ◽  
Domain Specific ◽  
Proposed Model ◽  
Segmentation Framework ◽  
Small Sample Problem

Clothing segmentation is a challenging vision problem typically implemented within a fine-grained semantic segmentation framework. Different from conventional segmentation, clothing segmentation has some domain-specific properties such as texture richness, diverse appearance variations, non-rigid geometry deformations, and small sample learning. To deal with these points, we propose a semantic locality-aware segmentation model, which adaptively attaches an original clothing image with a semantically similar (e.g., appearance or pose) auxiliary exemplar by search. Through considering the interactions of the clothing image and its exemplar, more intrinsic knowledge about the locality manifold structures of clothing images is discovered to make the learning process of small sample problem more stable and tractable. Furthermore, we present a CNN model based on the deformable convolutions to extract the non-rigid geometry-aware features for clothing images. Experimental results demonstrate the effectiveness of the proposed model against the state-of-the-art approaches.

scholarly journals A CNN Model for Human Parsing Based on Capacity Optimization

2019 ◽  
Vol 9 (7) ◽  
pp. 1330 ◽  
Author(s):  
Yalong Jiang ◽  
Zheru Chi
Keyword(s):  
Neural Networks ◽  
Computational Efficiency ◽  
Semantic Information ◽  
State Of The Art ◽  
Depth Estimation ◽  
Baseline Model ◽  
Computational Burden ◽  
Proposed Model ◽  
Saliency Prediction ◽  
Benchmark Solutions

Although a state-of-the-art performance has been achieved in pixel-specific tasks, such as saliency prediction and depth estimation, convolutional neural networks (CNNs) still perform unsatisfactorily in human parsing where semantic information of detailed regions needs to be perceived under the influences of variations in viewpoints, poses, and occlusions. In this paper, we propose to improve the robustness of human parsing modules by introducing a depth-estimation module. A novel scheme is proposed for the integration of a depth-estimation module and a human-parsing module. The robustness of the overall model is improved with the automatically obtained depth labels. As another major concern, the computational efficiency is also discussed. Our proposed human parsing module with 24 layers can achieve a similar performance as the baseline CNN model with over 100 layers. The number of parameters in the overall model is less than that in the baseline model. Furthermore, we propose to reduce the computational burden by replacing a conventional CNN layer with a stack of simplified sub-layers to further reduce the overall number of trainable parameters. Experimental results show that the integration of two modules contributes to the improvement of human parsing without additional human labeling. The proposed model outperforms the benchmark solutions and the capacity of our model is better matched to the complexity of the task.

scholarly journals Depthwise Convolution Is All You Need for Learning Multiple Visual Domains

2019 ◽  
Vol 33 ◽  
pp. 8368-8375 ◽  
Author(s):  
Yunhui Guo ◽  
Yandong Li ◽  
Liqiang Wang ◽  
Tajana Rosing
Keyword(s):  
State Of The Art ◽  
The State ◽  
Spatial Correlations ◽  
Single Model ◽  
Domain Specific ◽  
Gating Mechanism ◽  
Proposed Model ◽  
Domain Learning ◽  
Domain Models ◽  
Universal Structure

There is a growing interest in designing models that can deal with images from different visual domains. If there exists a universal structure in different visual domains that can be captured via a common parameterization, then we can use a single model for all domains rather than one model per domain. A model aware of the relationships between different domains can also be trained to work on new domains with less resources. However, to identify the reusable structure in a model is not easy. In this paper, we propose a multi-domain learning architecture based on depthwise separable convolution. The proposed approach is based on the assumption that images from different domains share cross-channel correlations but have domain-specific spatial correlations. The proposed model is compact and has minimal overhead when being applied to new domains. Additionally, we introduce a gating mechanism to promote soft sharing between different domains. We evaluate our approach on Visual Decathlon Challenge, a benchmark for testing the ability of multi-domain models. The experiments show that our approach can achieve the highest score while only requiring 50% of the parameters compared with the state-of-the-art approaches.

scholarly journals Improving the Quantification of DNA Sequences Using Evolutionary Information Based on Deep Learning

Cells ◽  
2019 ◽  
Vol 8 (12) ◽  
pp. 1635 ◽  
Author(s):  
Hilal Tayara ◽  
Kil To Chong
Keyword(s):  
Dna Sequences ◽  
Multiple Scales ◽  
State Of The Art ◽  
Characteristic Curve ◽  
The State ◽  
Evolutionary Information ◽  
Functional Variants ◽  
Proposed Model ◽  
Precision Recall Curve

It is known that over 98% of the human genome is non-coding, and 93% of disease associated variants are located in these regions. Therefore, understanding the function of these regions is important. However, this task is challenging as most of these regions are not well understood in terms of their functions. In this paper, we introduce a novel computational model based on deep neural networks, called DQDNN, for quantifying the function of non-coding DNA regions. This model combines convolution layers for capturing regularity motifs at multiple scales and recurrent layers for capturing long term dependencies between the captured motifs. In addition, we show that integrating evolutionary information with raw genomic sequences improves the performance of the predictor significantly. The proposed model outperforms the state-of-the-art ones using raw genomics sequences only and also by integrating evolutionary information with raw genomics sequences. More specifically, the proposed model improves 96.9% and 98% of the targets in terms of area under the receiver operating characteristic curve and the precision-recall curve, respectively. In addition, the proposed model improved the prioritization of functional variants of expression quantitative trait loci (eQTLs) compared with the state-of-the-art models.

scholarly journals SalSAC: A Video Saliency Prediction Model with Shuffled Attentions and Correlation-Based ConvLSTM

2020 ◽  
Vol 34 (07) ◽  
pp. 12410-12417 ◽  
Author(s):  
Xinyi Wu ◽  
Zhenyao Wu ◽  
Jinglin Zhang ◽  
Lili Ju ◽  
Song Wang
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Prediction Model ◽  
Convolutional Neural Networks ◽  
State Of The Art ◽  
Dynamic Aspect ◽  
Static Information ◽  
Saliency Prediction ◽  
Multi Level ◽  
Video Saliency

The performance of predicting human fixations in videos has been much enhanced with the help of development of the convolutional neural networks (CNN). In this paper, we propose a novel end-to-end neural network “SalSAC” for video saliency prediction, which uses the CNN-LSTM-Attention as the basic architecture and utilizes the information from both static and dynamic aspects. To better represent the static information of each frame, we first extract multi-level features of same size from different layers of the encoder CNN and calculate the corresponding multi-level attentions, then we randomly shuffle these attention maps among levels and multiply them to the extracted multi-level features respectively. Through this way, we leverage the attention consistency across different layers to improve the robustness of the network. On the dynamic aspect, we propose a correlation-based ConvLSTM to appropriately balance the influence of the current and preceding frames to the prediction. Experimental results on the DHF1K, Hollywood2 and UCF-sports datasets show that SalSAC outperforms many existing state-of-the-art methods.

scholarly journals MS-MDA: Multisource Marginal Distribution Adaptation for Cross-Subject and Cross-Session EEG Emotion Recognition

2021 ◽  
Vol 15 ◽  
Author(s):  
Hao Chen ◽  
Ming Jin ◽  
Zhunan Li ◽  
Cunhang Fan ◽  
Jinpeng Li ◽  
...  
Keyword(s):  
Emotion Recognition ◽  
Marginal Distribution ◽  
Domain Adaptation ◽  
State Of The Art ◽  
Essential Element ◽  
Data Share ◽  
Domain Specific ◽  
Eeg Data ◽  
Data Source ◽  
Electroencephalogram Eeg

As an essential element for the diagnosis and rehabilitation of psychiatric disorders, the electroencephalogram (EEG) based emotion recognition has achieved significant progress due to its high precision and reliability. However, one obstacle to practicality lies in the variability between subjects and sessions. Although several studies have adopted domain adaptation (DA) approaches to tackle this problem, most of them treat multiple EEG data from different subjects and sessions together as a single source domain for transfer, which either fails to satisfy the assumption of domain adaptation that the source has a certain marginal distribution, or increases the difficulty of adaptation. We therefore propose the multi-source marginal distribution adaptation (MS-MDA) for EEG emotion recognition, which takes both domain-invariant and domain-specific features into consideration. First, we assume that different EEG data share the same low-level features, then we construct independent branches for multiple EEG data source domains to adopt one-to-one domain adaptation and extract domain-specific features. Finally, the inference is made by multiple branches. We evaluate our method on SEED and SEED-IV for recognizing three and four emotions, respectively. Experimental results show that the MS-MDA outperforms the comparison methods and state-of-the-art models in cross-session and cross-subject transfer scenarios in our settings. Codes at https://github.com/VoiceBeer/MS-MDA.

scholarly journals Bi-Directional Generation for Unsupervised Domain Adaptation

2020 ◽  
Vol 34 (04) ◽  
pp. 6615-6622 ◽  
Author(s):  
Guanglei Yang ◽  
Haifeng Xia ◽  
Mingli Ding ◽  
Zhengming Ding
Keyword(s):  
Domain Adaptation ◽  
State Of The Art ◽  
Target Domain ◽  
Unsupervised Domain Adaptation ◽  
Cross Domain ◽  
Proposed Model ◽  
Latent Space ◽  
Domain Alignment ◽  
Domain Information ◽  
Adaptation Model

Unsupervised domain adaptation facilitates the unlabeled target domain relying on well-established source domain information. The conventional methods forcefully reducing the domain discrepancy in the latent space will result in the destruction of intrinsic data structure. To balance the mitigation of domain gap and the preservation of the inherent structure, we propose a Bi-Directional Generation domain adaptation model with consistent classifiers interpolating two intermediate domains to bridge source and target domains. Specifically, two cross-domain generators are employed to synthesize one domain conditioned on the other. The performance of our proposed method can be further enhanced by the consistent classifiers and the cross-domain alignment constraints. We also design two classifiers which are jointly optimized to maximize the consistency on target sample prediction. Extensive experiments verify that our proposed model outperforms the state-of-the-art on standard cross domain visual benchmarks.

Enhanced context-aware recommendation using topic modeling and particle swarm optimization

2021 ◽  
pp. 1-16
Author(s):  
Ibtissem Gasmi ◽  
Mohamed Walid Azizi ◽  
Hassina Seridi-Bouchelaghem ◽  
Nabiha Azizi ◽  
Samir Brahim Belhaouari
Keyword(s):  
Topic Modeling ◽  
Latent Dirichlet Allocation ◽  
State Of The Art ◽  
Weighting Function ◽  
Contextual Factors ◽  
Pearson Correlation ◽  
Correlation Coefficients ◽  
Pso Algorithm ◽  
Context Aware ◽  
Proposed Model

Context-Aware Recommender System (CARS) suggests more relevant services by adapting them to the user’s specific context situation. Nevertheless, the use of many contextual factors can increase data sparsity while few context parameters fail to introduce the contextual effects in recommendations. Moreover, several CARSs are based on similarity algorithms, such as cosine and Pearson correlation coefficients. These methods are not very effective in the sparse datasets. This paper presents a context-aware model to integrate contextual factors into prediction process when there are insufficient co-rated items. The proposed algorithm uses Latent Dirichlet Allocation (LDA) to learn the latent interests of users from the textual descriptions of items. Then, it integrates both the explicit contextual factors and their degree of importance in the prediction process by introducing a weighting function. Indeed, the PSO algorithm is employed to learn and optimize weights of these features. The results on the Movielens 1 M dataset show that the proposed model can achieve an F-measure of 45.51% with precision as 68.64%. Furthermore, the enhancement in MAE and RMSE can respectively reach 41.63% and 39.69% compared with the state-of-the-art techniques.

scholarly journals A Deep Learning Approach to Predict Autism Spectrum Disorder Using Multisite Resting-State fMRI

2021 ◽  
Vol 11 (8) ◽  
pp. 3636
Author(s):  
Faria Zarin Subah ◽  
Kaushik Deb ◽  
Pranab Kumar Dhar ◽  
Takeshi Koshiba
Keyword(s):  
Autism Spectrum Disorder ◽  
Resting State ◽  
State Of The Art ◽  
Resting State Fmri ◽  
Autism Spectrum ◽  
Spectrum Disorder ◽  
Bootstrap Analysis ◽  
Proposed Model ◽  
Art Methods ◽  
The Mean

Autism spectrum disorder (ASD) is a complex and degenerative neuro-developmental disorder. Most of the existing methods utilize functional magnetic resonance imaging (fMRI) to detect ASD with a very limited dataset which provides high accuracy but results in poor generalization. To overcome this limitation and to enhance the performance of the automated autism diagnosis model, in this paper, we propose an ASD detection model using functional connectivity features of resting-state fMRI data. Our proposed model utilizes two commonly used brain atlases, Craddock 200 (CC200) and Automated Anatomical Labelling (AAL), and two rarely used atlases Bootstrap Analysis of Stable Clusters (BASC) and Power. A deep neural network (DNN) classifier is used to perform the classification task. Simulation results indicate that the proposed model outperforms state-of-the-art methods in terms of accuracy. The mean accuracy of the proposed model was 88%, whereas the mean accuracy of the state-of-the-art methods ranged from 67% to 85%. The sensitivity, F1-score, and area under receiver operating characteristic curve (AUC) score of the proposed model were 90%, 87%, and 96%, respectively. Comparative analysis on various scoring strategies show the superiority of BASC atlas over other aforementioned atlases in classifying ASD and control.

Sign in / Sign up

Export Citation Format

Share Document