Lung pneumonia severity scoring in chest X-ray images using transformers.

To create robust and adaptable methods for lung pneumonia diagnosis and the assessment of its severity using chest X-rays (CXR), access to well-curated, extensive datasets is crucial. Many current severity quantification approaches require resource-intensive training for optimal results. Healthcare practitioners require efficient computational tools to swiftly identify COVID-19 cases and predict the severity of the condition. In this research, we introduce a novel image augmentation scheme as well as a neural network model founded on Vision Transformers (ViT) with a small number of trainable parameters for quantifying COVID-19 severity and other lung diseases. Our method, named Vision Transformer Regressor Infection Prediction (ViTReg-IP), leverages a ViT architecture and a regression head. To assess the model's adaptability, we evaluate its performance on diverse chest radiograph datasets from various open sources. We conduct a comparative analysis against several competing deep learning methods. Our results achieved a minimum Mean Absolute Error (MAE) of 0.569 and 0.512 and a maximum Pearson Correlation Coefficient (PC) of 0.923 and 0.855 for the geographic extent score and the lung opacity score, respectively, when the CXRs from the RALO dataset were used in training. The experimental results reveal that our model delivers exceptional performance in severity quantification while maintaining robust generalizability, all with relatively modest computational requirements. The source codes used in our work are publicly available at https://github.com/bouthainas/ViTReg-IP .

COVID-19 Infection Percentage Estimation from Computed Tomography Scans: Results and Insights from the International Per-COVID-19 Challenge.

COVID-19 analysis from medical imaging is an important task that has been intensively studied in the last years due to the spread of the COVID-19 pandemic. In fact, medical imaging has often been used as a complementary or main tool to recognize the infected persons. On the other hand, medical imaging has the ability to provide more details about COVID-19 infection, including its severity and spread, which makes it possible to evaluate the infection and follow-up the patient's state. CT scans are the most informative tool for COVID-19 infection, where the evaluation of COVID-19 infection is usually performed through infection segmentation. However, segmentation is a tedious task that requires much effort and time from expert radiologists. To deal with this limitation, an efficient framework for estimating COVID-19 infection as a regression task is proposed. The goal of the Per-COVID-19 challenge is to test the efficiency of modern deep learning methods on COVID-19 infection percentage estimation (CIPE) from CT scans. Participants had to develop an efficient deep learning approach that can learn from noisy data. In addition, participants had to cope with many challenges, including those related to COVID-19 infection complexity and crossdataset scenarios. This paper provides an overview of the COVID-19 infection percentage estimation challenge (Per-COVID-19) held at MIA-COVID-2022. Details of the competition data, challenges, and evaluation metrics are presented. The best performing approaches and their results are described and discussed.

LGCOAMix: Local and Global Context-and-Object-Part-Aware Superpixel-Based Data Augmentation for Deep Visual Recognition.

Cutmix-based data augmentation, which uses a cut-and-paste strategy, has shown remarkable generalization capabilities in deep learning. However, existing methods primarily consider global semantics with image-level constraints, which excessively reduces attention to the discriminative local context of the class and leads to a performance improvement bottleneck. Moreover, existing methods for generating augmented samples usually involve cutting and pasting rectangular or square regions, resulting in a loss of object part information. To mitigate the problem of inconsistency between the augmented image and the generated mixed label, existing methods usually require double forward propagation or rely on an external pre-trained network for object centering, which is inefficient. To overcome the above limitations, we propose LGCOAMix, an efficient context-aware and object-part-aware superpixel-based grid blending method for data augmentation. To the best of our knowledge, this is the first time that a label mixing strategy using a superpixel attention approach has been proposed for cutmix-based data augmentation. It is the first instance of learning local features from discriminative superpixel-wise regions and cross-image superpixel contrasts. Extensive experiments on various benchmark datasets show that LGCOAMix outperforms state-of-the-art cutmix-based data augmentation methods on classification tasks, and weakly supervised object location on CUB200-2011. We have demonstrated the effectiveness of LGCOAMix not only for CNN networks, but also for Transformer networks. Source codes are available at https://github.com/DanielaPlusPlus/LGCOAMix.

Boosted Additive Angular Margin Loss for breast cancer diagnosis from histopathological images.

Pathologists use biopsies and microscopic examination to accurately diagnose breast cancer. This process is time-consuming, labor-intensive, and costly. Convolutional neural networks (CNNs) offer an efficient and highly accurate approach to reduce analysis time and automate the diagnostic workflow in pathology. However, the softmax loss commonly used in existing CNNs leads to noticeable ambiguity in decision boundaries and lacks a clear constraint for minimizing within-class variance. In response to this problem, a solution in the form of softmax losses based on angular margin was developed. These losses were introduced in the context of face recognition, with the goal of integrating an angular margin into the softmax loss. This integration improves discrimination features during CNN training by effectively increasing the distance between different classes while reducing the variance within each class. Despite significant progress, these losses are limited to target classes only when margin penalties are applied, which may not lead to optimal effectiveness. In this paper, we introduce Boosted Additive Angular Margin Loss (BAM) to obtain highly discriminative features for breast cancer diagnosis from histopathological images. BAM not only penalizes the angle between deep features and their target class weights, but also considers angles between deep features and non-target class weights. We performed extensive experiments on the publicly available BreaKHis dataset. BAM achieved remarkable accuracies of 99.79%, 99.86%, 99.96%, and 97.65% for magnification levels of 40X, 100X, 200X, and 400X, respectively. These results show an improvement in accuracy of 0.13%, 0.34%, and 0.21% for 40X, 100X, and 200X magnifications, respectively, compared to the baseline methods. Additional experiments were performed on the BACH dataset for breast cancer classification and on the widely accepted LFW and YTF datasets for face recognition to evaluate the generalization ability of the proposed loss function. The results show that BAM outperforms state-of-the-art methods by increasing the decision space between classes and minimizing intra-class variance, resulting in improved discriminability.

A unified semi-supervised model with joint estimation of graph, soft labels and latent subspace.

Since manually labeling images is expensive and labor intensive, in practice we often do not have enough labeled images to train an effective classifier for the new image classification tasks. The graph-based SSL methods have received more attention in practice due to their convexity, scalability and efficiency. In this paper, we propose a novel graph-based semi-supervised learning method that takes full advantage of a small set of labeled graphs and a large set of unlabeled graph data. First, we explain the concept of graph-based semi-supervised learning. The core idea of these models is to jointly estimate a low-rank graph with soft labels and a latent subspace. The proposed scheme leverages the synergy between the graph structure and the data representation in terms of soft labels and latent features. This improves the monitoring information and leads to better discriminative linear transformation. Several experiments were conducted on five image datasets using state-of-the-art methods. These experiments show the effectiveness of the proposed semi-supervised method.

PDAtt-Unet: Pyramid Dual-Decoder Attention Unet for Covid-19 infection segmentation from CT-scans.

Since the emergence of the Covid-19 pandemic in late 2019, medical imaging has been widely used to analyze this disease. Indeed, CT-scans of the lungs can help diagnose, detect, and quantify Covid-19 infection. In this paper, we address the segmentation of Covid-19 infection from CT-scans. To improve the performance of the Att-Unet architecture and maximize the use of the Attention Gate, we propose the PAtt-Unet and DAtt-Unet architectures. PAtt-Unet aims to exploit the input pyramids to preserve the spatial awareness in all of the encoder layers. On the other hand, DAtt-Unet is designed to guide the segmentation of Covid-19 infection inside the lung lobes. We also propose to combine these two architectures into a single one, which we refer to as PDAtt-Unet. To overcome the blurry boundary pixels segmentation of Covid-19 infection, we propose a hybrid loss function. The proposed architectures were tested on four datasets with two evaluation scenarios (intra and cross datasets). Experimental results showed that both PAtt-Unet and DAtt-Unet improve the performance of Att-Unet in segmenting Covid-19 infections. Moreover, the combination architecture PDAtt-Unet led to further improvement. To Compare with other methods, three baseline segmentation architectures (Unet, Unet++, and Att-Unet) and three state-of-the-art architectures (InfNet, SCOATNet, and nCoVSegNet) were tested. The comparison showed the superiority of the proposed PDAtt-Unet trained with the proposed hybrid loss (PDEAtt-Unet) over all other methods. Moreover, PDEAtt-Unet is able to overcome various challenges in segmenting Covid-19 infections in four datasets and two evaluation scenarios.

A unified deep semi-supervised graph learning scheme based on nodes re-weighting and manifold regularization.

In recent years, semi-supervised learning on graphs has gained importance in many fields and applications. The goal is to use both partially labeled data (labeled examples) and a large amount of unlabeled data to build more effective predictive models. Deep Graph Neural Networks (GNNs) are very useful in both unsupervised and semi-supervised learning problems. As a special class of GNNs, Graph Convolutional Networks (GCNs) aim to obtain data representation through graph-based node smoothing and layer-wise neural network transformations. However, GCNs have some weaknesses when applied to semi-supervised graph learning: (1) it ignores the manifold structure implicitly encoded by the graph; (2) it uses a fixed neighborhood graph and focuses only on the convolution of a graph, but pays little attention to graph construction; (3) it rarely considers the problem of topological imbalance. To overcome the above shortcomings, in this paper, we propose a novel semi-supervised learning method called Re-weight Nodes and Graph Learning Convolutional Network with Manifold Regularization (ReNode-GLCNMR). Our proposed method simultaneously integrates graph learning and graph convolution into a unified network architecture, which also enforces label smoothing through an unsupervised loss term. At the same time, it addresses the problem of imbalance in graph topology by adaptively reweighting the influence of labeled nodes based on their distances to the class boundaries. Experiments on 8 benchmark datasets show that ReNode-GLCNMR significantly outperforms the state-of-the-art semi-supervised GNN methods.1.

Deep Learning with Discriminative Margin Loss for Cross-Domain Consumer-to-Shop Clothes Retrieval.

Consumer-to-shop clothes retrieval refers to the problem of matching photos taken by customers with their counterparts in the shop. Due to some problems, such as a large number of clothing categories, different appearances of clothing items due to different camera angles and shooting conditions, different background environments, and different body postures, the retrieval accuracy of traditional consumer-to-shop models is always low. With advances in convolutional neural networks (CNNs), the accuracy of garment retrieval has been significantly improved. Most approaches addressing this problem use single CNNs in conjunction with a softmax loss function to extract discriminative features. In the fashion domain, negative pairs can have small or large visual differences that make it difficult to minimize intraclass variance and maximize interclass variance with softmax. Margin-based softmax losses such as Additive Margin-Softmax (aka CosFace) improve the discriminative power of the original softmax loss, but since they consider the same margin for the positive and negative pairs, they are not suitable for cross-domain fashion search. In this work, we introduce the cross-domain discriminative margin loss (DML) to deal with the large variability of negative pairs in fashion. DML learns two different margins for positive and negative pairs such that the negative margin is larger than the positive margin, which provides stronger intraclass reduction for negative pairs. The experiments conducted on publicly available fashion datasets DARN and two benchmarks of the DeepFashion dataset-(1) Consumer-to-Shop Clothes Retrieval and (2) InShop Clothes Retrieval-confirm that the proposed loss function not only outperforms the existing loss functions but also achieves the best performance.

A Comparative Analysis of Human Behavior Prediction Approaches in Intelligent Environments.

Behavior modeling has multiple applications in the intelligent environment domain. It has been used in different tasks, such as the stratification of different pathologies, prediction of the user actions and activities, or modeling the energy usage. Specifically, behavior prediction can be used to forecast the future evolution of the users and to identify those behaviors that deviate from the expected conduct. In this paper, we propose the use of embeddings to represent the user actions, and study and compare several behavior prediction approaches. We test multiple model (LSTM, CNNs, GCNs, and transformers) architectures to ascertain the best approach to using embeddings for behavior modeling and also evaluate multiple embedding retrofitting approaches. To do so, we use the Kasteren dataset for intelligent environments, which is one of the most widely used datasets in the areas of activity recognition and behavior modeling.

SuperpixelGridMasks Data Augmentation: Application to Precision Health and Other Real-world Data.

A novel approach of data augmentation based on irregular superpixel decomposition is proposed. This approach called SuperpixelGridMasks permits to extend original image datasets that are required by training stages of machine learning-related analysis architectures towards increasing their performances. Three variants named SuperpixelGridCut, SuperpixelGridMean, and SuperpixelGridMix are presented. These grid-based methods produce a new style of image transformations using the dropping and fusing of information. Extensive experiments using various image classification models as well as precision health and surrounding real-world datasets show that baseline performances can be significantly outperformed using our methods. The comparative study also shows that our methods can overpass the performances of other data augmentations. SuperpixelGridCut, SuperpixelGridMean, and SuperpixelGridMix codes are publicly available at https://github.com/hammoudiproject/SuperpixelGridMasks.

Multiple-view flexible semi-supervised classification through consistent graph construction and label propagation.

Graph construction plays an essential role in graph-based label propagation since graphs give some information on the structure of the data manifold. While most graph construction methods rely on predefined distance calculation, recent algorithms merge the task of label propagation and graph construction in a single process. Moreover, the use of several descriptors is proved to outperform a single descriptor in representing the relation between the nodes. In this article, we propose a Multiple-View Consistent Graph construction and Label propagation algorithm (MVCGL) that simultaneously constructs a consistent graph based on several descriptors and performs label propagation over unlabeled samples. Furthermore, it provides a mapping function from the feature space to the label space with which we estimate the label of unseen samples via a linear projection. The constructed graph does not rely on a predefined similarity function and exploits data and label smoothness. Experiments conducted on three face and one handwritten digit databases show that the proposed method can gain better performance compared to other graph construction and label propagation methods.

Joint Label Inference and Discriminant Embedding.

Graph-based learning in semisupervised models provides an effective tool for modeling big data sets in high-dimensional spaces. It has been useful for propagating a small set of initial labels to a large set of unlabeled data. Thus, it meets the requirements of many emerging applications. However, in real-world applications, the scarcity of labeled data can negatively affect the performance of the semisupervised method. In this article, we present a new framework for semisupervised learning called joint label inference and discriminant embedding for soft label inference and linear feature extraction. The proposed criterion and its associated optimization algorithm take advantage of both labeled and unlabeled data samples in order to estimate the discriminant transformation. This type of criterion should allow learning more discriminant semisupervised models. Nine public image data sets are used in the experiments and method comparisons. These experimental results show that the performance of the proposed method is superior to that of many advanced semisupervised graph-based algorithms.

Per-COVID-19: A Benchmark Dataset for COVID-19 Percentage Estimation from CT-Scans.

COVID-19 infection recognition is a very important step in the fight against the COVID-19 pandemic. In fact, many methods have been used to recognize COVID-19 infection including Reverse Transcription Polymerase Chain Reaction (RT-PCR), X-ray scan, and Computed Tomography scan (CT- scan). In addition to the recognition of the COVID-19 infection, CT scans can provide more important information about the evolution of this disease and its severity. With the extensive number of COVID-19 infections, estimating the COVID-19 percentage can help the intensive care to free up the resuscitation beds for the critical cases and follow other protocol for less severity cases. In this paper, we introduce COVID-19 percentage estimation dataset from CT-scans, where the labeling process was accomplished by two expert radiologists. Moreover, we evaluate the performance of three Convolutional Neural Network (CNN) architectures: ResneXt-50, Densenet-161, and Inception-v3. For the three CNN architectures, we use two loss functions: MSE and Dynamic Huber. In addition, two pretrained scenarios are investigated (ImageNet pretrained models and pretrained models using X-ray data). The evaluated approaches achieved promising results on the estimation of COVID-19 infection. Inception-v3 using Dynamic Huber loss function and pretrained models using X-ray data achieved the best performance for slice-level results: 0.9365, 5.10, and 9.25 for Pearson Correlation coefficient (PC), Mean Absolute Error (MAE), and Root Mean Square Error (RMSE), respectively. On the other hand, the same approach achieved 0.9603, 4.01, and 6.79 for PCsubj, MAEsubj, and RMSEsubj, respectively, for subject-level results. These results prove that using CNN architectures can provide accurate and fast solution to estimate the COVID-19 infection percentage for monitoring the evolution of the patient state.

Deep Learning on Chest X-ray Images to Detect and Evaluate Pneumonia Cases at the Era of COVID-19.

Coronavirus disease 2019 (COVID-19) is an infectious disease with first symptoms similar to the flu. COVID-19 appeared first in China and very quickly spreads to the rest of the world, causing then the 2019-20 coronavirus pandemic. In many cases, this disease causes pneumonia. Since pulmonary infections can be observed through radiography images, this paper investigates deep learning methods for automatically analyzing query chest X-ray images with the hope to bring precision tools to health professionals towards screening the COVID-19 and diagnosing confirmed patients. In this context, training datasets, deep learning architectures and analysis strategies have been experimented from publicly open sets of chest X-ray images. Tailored deep learning models are proposed to detect pneumonia infection cases, notably viral cases. It is assumed that viral pneumonia cases detected during an epidemic COVID-19 context have a high probability to presume COVID-19 infections. Moreover, easy-to-apply health indicators are proposed for estimating infection status and predicting patient status from the detected pneumonia cases. Experimental results show possibilities of training deep learning models over publicly open sets of chest X-ray images towards screening viral pneumonia. Chest X-ray test images of COVID-19 infected patients are successfully diagnosed through detection models retained for their performances. The efficiency of proposed health indicators is highlighted through simulated scenarios of patients presenting infections and health problems by combining real and synthetic health data.

Learning a discriminant graph-based embedding with feature selection for image categorization.

Graph-based embedding methods are very useful for reducing the dimension of high-dimensional data and for extracting their relevant features. In this paper, we introduce a novel nonlinear method called Flexible Discriminant graph-based Embedding with feature selection (FDEFS). The proposed algorithm aims to classify image sample data in supervised learning and semi-supervised learning settings. Specifically, our method incorporates the Manifold Smoothness, Margin Discriminant Embedding and the Sparse Regression for feature selection. The weights add ℓ2,1-norm regularization for local linear approximation. The sparse regression implicitly performs feature selection on the original features of data matrix and of the linear transform. We also provide an effective solution method to optimize the objective function. We apply the algorithm on six public image datasets including scene, face and object datasets. These experiments demonstrate the effectiveness of the proposed embedding method. They also show that proposed the method compares favorably with many competing embedding methods.

Graph construction using adaptive Local Hybrid Coding scheme.

It is well known that dense coding with local bases (via Least Square coding schemes) can lead to large quantization errors or poor performances of machine learning tasks. On the other hand, sparse coding focuses on accurate representation without taking into account data locality due to its tendency to ignore the intrinsic structure hidden among the data. Local Hybrid Coding (LHC) (Xiang et al., 2014) was recently proposed as an alternative to the sparse coding scheme that is used in Sparse Representation Classifier (SRC). The LHC blends sparsity and bases-locality criteria in a unified optimization problem. It can retain the strengths of both sparsity and locality. Thus, the hybrid codes would have some advantages over both dense and sparse codes. This paper introduces a data-driven graph construction method that exploits and extends the LHC scheme. In particular, we propose a new coding scheme coined Adaptive Local Hybrid Coding (ALHC). The main contributions are as follows. First, the proposed coding scheme adaptively selects the local and non-local bases of LHC using data similarities provided by Locality-constrained Linear code. Second, the proposed ALHC exploits local similarities in its solution. Third, we use the proposed coding scheme for graph construction. For the task of graph-based label propagation, we demonstrate high classification performance of the proposed graph method on four benchmark face datasets: Extended Yale, PF01, PIE, and FERET.

Learning Flexible Graph-Based Semi-Supervised Embedding.

This paper introduces a graph-based semi-supervised embedding method as well as its kernelized version for generic classification and recognition tasks. The aim is to combine the merits of flexible manifold embedding and nonlinear graph-based embedding for semi-supervised learning. The proposed linear method will be flexible since it estimates a nonlinear manifold that is the closest one to a linear embedding. The proposed kernelized method will also be flexible since it estimates a kernel-based embedding that is the closest to a nonlinear manifold. In both proposed methods, the nonlinear manifold and the mapping (linear transform for the linear method and the kernel multipliers for the kernelized method) are simultaneously estimated, which overcomes the shortcomings of a cascaded estimation. The dimension of the final embedding obtained by the two proposed methods is not limited to the number of classes. They can be used by any kind of classifiers once the data are embedded into the new subspaces. Unlike nonlinear dimensionality reduction approaches, which suffer from out-of-sample problem, our proposed methods have an obvious advantage that the learnt subspace has a direct out-of-sample extension to novel samples, and are thus easily generalized to the entire high-dimensional input space. We provide extensive experiments on seven public databases in order to study the performance of the proposed methods. These experiments demonstrate much improvement over the state-of-the-art algorithms that are based on label propagation or graph-based semi-supervised embedding.

Exponential local discriminant embedding and its application to face recognition.

Local discriminant embedding (LDE) has been recently proposed to overcome some limitations of the global linear discriminant analysis method. In the case of a small training data set, however, LDE cannot directly be applied to high-dimensional data. This case is the so-called small-sample-size (SSS) problem. The classical solution to this problem was applying dimensionality reduction on the raw data (e.g., using principal component analysis). In this paper, we introduce a novel discriminant technique called "exponential LDE" (ELDE). The proposed ELDE can be seen as an extension of LDE framework in two directions. First, the proposed framework overcomes the SSS problem without discarding the discriminant information that was contained in the null space of the locality preserving scatter matrices associated with LDE. Second, the proposed ELDE is equivalent to transforming original data into a new space by distance diffusion mapping (similar to kernel-based nonlinear mapping), and then, LDE is applied in such a new space. As a result of diffusion mapping, the margin between samples belonging to different classes is enlarged, which is helpful in improving classification accuracy. The experiments are conducted on five public face databases: Yale, Extended Yale, PF01, Pose, Illumination, and Expression (PIE), and Facial Recognition Technology (FERET). The results show that the performances of the proposed ELDE are better than those of LDE and many state-of-the-art discriminant analysis techniques.

Efficient object detection and tracking in video sequences.

One of the most important problems in computer vision is the computation of the two-dimensional projective transformation (homography) that maps features of planar objects in different images and videos. This computation is required by many applications such as image mosaicking, image registration, and augmented reality. The real-time performance imposes constraints on the methods used. In this paper, we address the real-time detection and tracking of planar objects in a video sequence where the object of interest is given by a reference image template. Most existing approaches for homography estimation are based on two steps: feature extraction (first step) followed by a combinatorial optimization method (second step) to match features between the reference template and the scene frame. This paper has two main contributions. First, we detect both planar and nonplanar objects via efficient object feature classification in the input images, which is applied prior to performing the matching step. Second, for the tracking part (planar objects), we propose a fast method for the computation of the homography that is based on the transferred object features and their associated local raw brightness. The advantage of the proposed schemes is a fast matching as well as fast and robust object registration that is given by either a homography or three-dimensional pose.

A featureless approach to 3D polyhedral building modeling from aerial images.

This paper presents a model-based approach for reconstructing 3D polyhedral building models from aerial images. The proposed approach exploits some geometric and photometric properties resulting from the perspective projection of planar structures. Data are provided by calibrated aerial images. The novelty of the approach lies in its featurelessness and in its use of direct optimization based on image rawbrightness. The proposed framework avoids feature extraction and matching. The 3D polyhedral model is directly estimated by optimizing an objective function that combines an image-based dissimilarity measure and a gradient score over several aerial images. The optimization process is carried out by the Differential Evolution algorithm. The proposed approach is intended to provide more accurate 3D reconstruction than feature-based approaches. Fast 3D model rectification and updating can take advantage of the proposed method. Several results and evaluations of performance from real and synthetic images show the feasibility and robustness of the proposed approach.