Abstract
PurposeFresh food products remain one of the most challenging product categories for e-commerce managers. The literature emphasizes the importance of perceived freshness in explaining their purchase behavior. However, studies on online purchases of fresh food products are scarce, especially regarding repurchase intentions, and the role of perceived freshness in online settings has so far been disregarded. This research addresses this gap by examining the role of perceived freshness in the intention to repurchase fresh food products online.Design/methodology/approachGuided by the expectation confirmation theory (ECT) and the perceived risk theory, this study defined a set of hypotheses tested through structural equation modeling. Participants were consumers with previous experience in purchasing fresh food products online.FindingsThe findings indicate that the importance of sensory attributes negatively affected the perceived freshness of fresh food products purchased online, while the importance of non-sensory attributes had a non-significant impact. Expectations of freshness positively affected perceived freshness and confirmation of freshness, as suggested by ECT. The hypothesized positive effects of confirmation on satisfaction and of satisfaction on intention to repurchase fresh food products online were also supported. Finally, it was found that repurchase intention was negatively affected by perceived performance risk and financial risk.Originality/valueThis article contributes to the limited literature on online purchase of fresh food by focusing on perceived freshness as a determinant of repurchase intention.

Abstract
COVID-19 is an extremely contagious respiratory sickness instigated by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Common symptoms encompass fever, cough, fatigue, and breathing difficulties, often leading to hospitalization and fatalities in severe cases. CTCovid19 is a novel model tailored for COVID-19 detection, specifically honing in on a distinct deep learning structure, ResNet-50 trained with ImageNet serves as the foundational framework for our model. To enhance its capability to capture pertinent features related to COVID-19 patterns in Computed Tomography scans, the network underwent fine-tuning through layer adjustments and the addition of new ones. The model achieved accuracy rates that went from 97.0 % to 99.8 % across three widely recognized and documented datasets dedicated to COVID-19 detection. © 2024 The Authors

Abstract
Long-tailed data are a special type of multiclass imbalanced data with a very large amount of minority/tail classes that have a very significant combined influence. Long-tailed learning (LTL) aims to build high-performance models on datasets with long-tailed distributions that can identify all the classes with high accuracy, in particular the minority/tail classes. It is a cutting-edge research direction that has attracted a remarkable amount of research effort in the past few years. In this article, we present a comprehensive survey of the latest advances in long-tailed visual learning. We first propose a new taxonomy for LTL, which consists of eight different dimensions, including data balancing, neural architecture, feature enrichment, logits adjustment, loss function, bells and whistles, network optimization, and posthoc processing techniques. Based on our proposed taxonomy, we present a systematic review of LTL methods, discussing their commonalities and alignable differences. We also analyze the differences between imbalance learning and LTL. Finally, we discuss prospects and future directions in this field.

Abstract
Planar surfaces are commonly found in man-made underwater environments and can be employed to support underwater SLAM. This work focuses on 3D plane extraction, building on two-dimensional acoustic scans collected from an imaging sonar. The novel contribution of our algorithm exploits the sonar's wider beamwidth and ability to collect secondary echoes from these structures to extract a three-dimensional surface from the acquired acoustic image. Building on a Hough Transform-based algorithm adapted to polar-based acoustic imagery, line feature detection supports plane representation segmentation. An inverse sensor model is subsequently employed to estimate additional plane parameters: inclination, length, and height. Experimental assessment in a confined controlled environment is introduced, validating the accuracy of the algorithm. Additional results from a dam shaft scenario are also presented to assess the potential of the developed tool.

Abstract
Deep learning has been extensively applied to medical imaging tasks over the past years, achieving outstanding results. However, the obscure reasoning of the models and the lack of supportive evidence causes both clinicians and patients to distrust the models' predictions, hindering their adoption in clinical practice. In recent years, the research community has focused on developing explanations capable of revealing a model's reasoning. Among various types of explanations, example-based explanations emerged as particularly intuitive for medical practitioners. Despite the intuitiveness and wide development of example-based explanations, no work provides a comprehensive review of existing example-based explainability works in the medical image domain. In this work, we review works that provide example-based explanations for medical imaging tasks, reflecting on their strengths and limitations. We identify the absence of objective evaluation metrics, the lack of clinical validation and privacy concerns as the main issues that hinder the deployment of example-based explanations in clinical practice. Finally, we reflect on future directions contributing towards the deployment of example-based explainability in clinical practice.

Abstract
We propose a hybrid formal verification approach that combines high-level deductive reasoning and circuit-based reasoning and apply it to highly optimized cryptographic assembly code. Our approach permits scaling up formal verification in two complementary directions: 1) it reduces the proof effort required for low-level functions where the computation logics are obfuscated by the intricate use of architecture-specific instructions and 2) it permits amortizing the effort of proving one implementation by using equivalence checking to propagate the guarantees to other implementations of the same computation using different optimizations or targeting different architectures. We demonstrate our approach via an extension to the EasyCrypt proof assistant and by revisiting formally verified implementations of ML-KEM in Jasmin. As a result, we obtain the first formally verified implementation of ML-KEM that offers performance comparable to the fastest non-verified implementation in x86-64 architectures.