Deep learning models enhance breast cancer detection in mammograms but struggle with domain shifts, where test data differ from training data. Domain adaptation (DA) helps address this issue but often relies on unstable adversarial techniques. Breast lesion classification in mammograms also faces challenges like data scarcity and overfitting. Mixup mitigates these by generating synthetic samples, increasing variability, and improving robustness. Meanwhile, contrastive learning enhances feature alignment, boosting generalization and classification accuracy across domains. This paper proposes a DA model that integrates mixup and contrastive learning to improve feature alignment and generalization, leading to more accurate breast lesion classification. Our approach outperforms standard DA methods, achieving 82.5% accuracy, 0.774 F1 score, and 0.7868 AUC on INbreast (target dataset), surpassing DANN (63.6%) and Deep CORAL (67.7%). It also generalizes well, reaching 70.4% accuracy on CMMD and 63.64% on CDD-CESM, demonstrating its effectiveness in addressing domain shifts.

The performance of diagnostic Computer-Aided Design (CAD) systems for retinal diseases depends on the quality of the retinal images being screened. Thus, many studies have been developed to evaluate and assess the quality of such retinal images. However, most of them did not investigate the relationship between the accuracy of the developed models and the quality of the visualization of interpretability methods for distinguishing between gradable and non-gradable retinal images. Consequently, this paper presents a novel framework called “FGR-Net” to automatically asses and interpret underlying fundus image quality by merging an autoencoder network with a classifier network. The FGR-Net model also provides an interpretable quality assessment through visualizations. In particular, FGR-Net uses a deep autoencoder to reconstruct the input image in order to extract the visual characteristics of the input fundus images based on self-supervised learning. The extracted features by the autoencoder are then fed into a deep classifier network to distinguish between gradable and ungradable fundus images. FGR-Net is evaluated with different interpretability methods, which indicates that the autoencoder is a key factor in forcing the classifier to focus on the relevant structures of the fundus images, such as the fovea, optic disc, and prominent blood vessels. Additionally, the interpretability methods can provide visual feedback for ophthalmologists to understand how our model evaluates the quality of fundus images. The experimental results showed the superiority of FGR-Net over the state-of-the-art quality assessment methods, with an accuracy of >89% and an F1-score of >87%. The code is publicly available at https://github.com/saifalkh/FGR-Net.

Abstract Predicting human motion based on past observed motion is one of the challenging issues in computer vision and graphics. Existing research works are dealing with this issue by using discriminative models and showing the results for cases that follow a homogeneous distribution (in distribution) and not discussing the issues of the domain shift problem, where training and testing data follow a heterogeneous (out of distribution) problem, which is the reality when such models are used in practice. However, recent research proposed addressing domain shift issues by augmenting the discriminative model with a generative model and obtained better results. In the present investigation, we propose regularizing the extended network by inserting linear layers to minimize the rank of the latent space and train the entire end-to-end network. We regularize the network to strengthen the model to deal effectively with domain shift scenarios. Both training and testing data come from different distribution sets; to deal with this, we toughen our network by adding the extra linear layers to the network encoder. We tested our model with the benchmark datasets, CMU Motion Capture and Human3.6M, and proved that our model outperforms 14 OoD actions of H3.6M and 7 OoD actions of CMU MoCap in terms of the Euclidean distance calculated between predicted and ground truth joint angle values. Our average results of 14 OoD actions for short-term (80, 160, 320, 400) are 0.34, 0.6, 0.96, 1.07, and for CMU MoCap of 7 OoD actions for short-term and long term (80, 160, 320, 400, 1000) are 0.28, 0.45, 0.77, 0.89, 1.46. All these results are much better than the other state-of-the-art results.

This paper proposes an effective deep learning-based model for crack detection in images acquired by different acquisition systems (e.g., cameras mounted on vehicles and drones and smartphone cameras) in six countries. By utilizing successful training procedures and including multi-scale feature extraction models, the ensemble model is built using effective variations of the cutting-edge object detection technique, Yolov7. The top crack detection models are fused using the non-maximum suppression method to create the proposed ensemble model. The proposed crack detection model is trained and validated using the crowdsensing-based road damage detection challenge (CRDDC2022). With the test set, the proposed model produced an average F1 score of 0.65 with all leaderboards of CRDDC2022 (all countries, India, Japan, Norway, and the United States leaderboards). Our approach is ranked in the 5thposition in the CRDDC2022 challenge1. The source code is available at https://github.com/AmmarOkran/CRDD2022.

The determination of precise skin lesion boundaries in dermoscopic images using automated methods faces many challenges, most importantly, the presence of hair, inconspicuous lesion edges and low contrast in dermoscopic images, and variability in the color, texture and shapes of skin lesions. Existing deep learning-based skin lesion segmentation algorithms are expensive in terms of computational time and memory. Consequently, running such segmentation algorithms requires a powerful GPU and high bandwidth memory, which are not available in dermoscopy devices. Thus, this article aims to achieve precise skin lesion segmentation with minimum resources: a lightweight, efficient generative adversarial network (GAN) model called SLSNet, which combines 1-D kernel factorized networks, position and channel attention, and multiscale aggregation mechanisms with a GAN model. The 1-D kernel factorized network reduces the computational cost of 2D filtering. The position and channel attention modules enhance the discriminative ability between the lesion and non-lesion feature representations in spatial and channel dimensions, respectively. A multiscale block is also used to aggregate the coarse-to-fine features of input skin images and reduce the effect of the artifacts. SLSNet is evaluated on two publicly available datasets: ISBI 2017 and the ISIC 2018. Although SLSNet has only 2.35 million parameters, the experimental results demonstrate that it achieves segmentation results on a par with the state-of-the-art skin lesion segmentation methods with an accuracy of 97.61%, and Dice and Jaccard similarity coefficients of 90.63% and 81.98%, respectively. SLSNet can run at more than 110 frames per second (FPS) in a single GTX1080Ti GPU, which is faster than well-known deep learning-based image segmentation models, such as FCN. Therefore, SLSNet can be used for practical dermoscopic applications.

The intermittent nature associated with photovoltaic (PV) generation is a challenging problem for the optimal planning and efficient management in smart grids. A reliable forecasting model of solar irradiance can play an essential role in allowing high PV penetrations without degrading the grid performance. For this purpose, most related works either use individual forecasting models or ensemble approaches (e.g., weighted average), ignoring the interaction between the values to be aggregated and thus may worsen the forecasting reliability. Differently, in this article, we propose a reliable solar irradiance forecasting method based on long short-term memory (LSTM) models and an aggregation function based on Choquet integral. This novel combination has the following features: 1) LSTM models can achieve accurate predictions because they model the temporal changes in solar irradiance, thanks to their recurrent architecture and memory units, and 2) the Choquet integral can model the interaction between the inputs to be aggregated through a fuzzy measure. This aggregation technique can determine the largest consistency among the conflicting forecasting results, taking advantage of each individual model. To demonstrate the effectiveness of the proposed approach, we compare it with several forecasting methods using six realistic datasets collected from different sites in Finland in which solar irradiance is intermittent. The comparison reveals the high reliability of the proposed forecasting model with different sites and solar profiles.

In this paper we analyse the performance of various texture analysis methods for the purpose of reducing the number of false positives in breast cancer detection; as a result, the cost of breast cancer diagnosis would be reduced. We consider well-known methods such as local binary patterns, histogram of oriented gradients, co-occurrence matrix features and Gabor filters. Moreover, we propose the use of local directional number patterns as a new feature extraction method for breast mass detection. For each method, different classifiers are trained on the extracted features to predict the class of unknown instances. In order to improve the mass detection capability of each individual method, we use feature combination techniques and classifier majority voting. Some experiments were performed on the images obtained from a public breast cancer database, achieving promising levels of sensitivity and specificity.