Mid Sweden University

Recognising the growth and transformation of plants over time is of great significance for both agriculture and environmental preservation. It aids stakeholders in efficiently utilising resources and providing optimal growth conditions. This paper investigates the automated Plant Growth Temporal Labelling (PGTempLab) problem, which involves plant growth classification based on visual imagery—specifically focusing on pine seedlings, a key forest species in Scandinavia. A central challenge of PGTempLab lies in the Class-Based Variability (CBV) Paradox, where significant intra-class variability and inter-class similarity occur due to temporal categorisation, complicating traditional classification tasks. To address this issue, a transfer learning approach is adopted using 12 well-known convolutional neural networks (CNNs) as pre-trained models. Additionally, three novel evaluation criteria are introduced, incorporating temporal aspects: time-dependent accuracy (TDA), temporal boundary deviation (TBD), and a temporal confusion matrix. Experiments, conducted on a custom dataset collected specifically for this study, consisting of over 38,000 time-lapse images of pine seedlings captured from two camera perspectives over 44 days, demonstrate that MobileNet achieves the highest performance in the top-view setting with 86.85% TDA, while SqueezeNetV2 performs best in the angle-view setting with 86.94% TDA. These results highlight the effectiveness of deep transfer learning in addressing the CBV paradox in plant growth classification. 

Detecting advertisements in digitized newspapers is a key step in large-scale media analytics and digital archiving. However, variations in layout, typography, and advertisement design across publishers and time periods cause significant domain shifts that reduce the generalization ability of supervised detectors. This paper presents AdAPT, a confidence-guided pseudo-labeling pipeline for unsupervised domain adaptation in advertisement detection. The proposed method leverages both advertisement-free (Null) and advertisement-containing pages from unlabeled target domains to generate reliable pseudo-labels. By retraining a YOLO-based detector using labeled source data combined with filtered pseudo-labeled target samples, AdAPT achieves robust adaptation without requiring manual annotation. Experiments conducted on two unseen newspapers (Adresseavisen and iTromsø) demonstrate that Null-based pseudo-labeling provides the most stable and accurate adaptation, yielding up to 38% error reduction compared to the baseline. The results highlight AdAPT as a simple, scalable, and annotation-efficient solution for maintaining high-performance advertisement detection across diverse newspaper collections. 

Deploying deep neural networks (DNNs) in real-time, resource-constrained edge environments remains a significant challenge due to the increasing depth and computational demand of state-of-the-art models. While state-of-the-art split computing and early-exit architectures independently offer solutions to reduce inference latency and offloading overhead, they are rarely integrated in a unified framework. This paper introduces BranchySplit, a novel architecture that combines runtime-adaptable state-of-the-art split computing with early exit mechanisms to accelerate inference in edge-cloud systems. Our method employs a shared runtime-adaptable autoencoder to support multiple partitioning points, enabling runtime adaptability based on varying system conditions. By integrating early exit branches at intermediate layers, the model can terminate inference early when confidence is sufficient, further reducing computation and communication. Extensive experiments on Caltech-256 and Tiny-ImageNet datasets using four backbone models (VGG-16, ResNet-50, EfficientFormerV2-S2, and ConvNeXT-Small) show that our approach consistently outperforms the traditional split computing approaches, achieving up to 2.7x latency speed-up compared to an all-in-server solution, while maintaining comparable or even improved accuracy. In addition, our proposed runtime-adaptable compression pipeline achieves a higher speed-up compared to the pipeline in similar work, i.e. SPINN, with up to 10x for ResNet50 on Caltech-256. These results demonstrate the efficacy of BranchySplit in enabling low-latency, resource-efficient vision inference in heterogeneous edge-cloud settings. 

The efficient deployment of deep neural networks (DNNs) on resource-constrained edge devices necessitates a delicate balance between model accuracy, communication bandwidth, and computational latency. While existing methods often address partitioning, quantization, or compression in isolation, this paper introduces CO-NAS (Joint Neural Architecture Search for Compression, Quantization, and Edge-Cloud Model Offloading), a unified framework that jointly optimizes these three critical dimensions for collaborative edge-cloud inference. By utilizing an adapted version of the stochastic architecture search (SNAS) paradigm, CO-NAS leverages Temperature-Annealed Softmax relaxations to enable gradient-based optimization over discrete choices of partition points, compression ratios, and activation bit-widths. Extensive experiments on ResNet-50 and VGG-16 across CIFAR-100 and Fashion-MNIST datasets demonstrate the framework's efficacy. CO-NAS discovers Pareto-optimal configurations, such as a ResNet-50 model for CIFAR-100 achieving 70.1% accuracy with extreme compression (48× channel reduction and 1-bit quantization), representing a total compression of over 1500× on transmitted features. Crucially, exhaustive grid search validation confirms that CO-NAS reliably identifies top-performing architectures within the vast search space, while achieving a 21× -27× reduction in search cost compared to the exhaustive baseline. The results validate CO-NAS as a scalable and efficient approach for automating the co-design of accurate and communication-efficient neural networks for the edge-cloud paradigm. 

Magnetic Resonance Imaging (MRI) is central to the diagnosis of neurological diseases such as Alzheimer’s disease (AD) and brain tumours, where early and reliable classification remains a pressing clinical need. In recent years, deep learning has transformed medical image analysis, with convolutional neural networks (CNNs), transformers, and hybrid models achieving unprecedented accuracy. However, despite these advances, challenges persist in areas such as generalisation across datasets, interpretability of model decisions, robustness to domain shifts, and the integration of multi-modal information. This paper makes two main contributions. First, we provide a comprehensive survey of deep learning methods for MRI-based classification of AD and brain tumours, covering CNNs, transformer-based models, hybrid designs, transfer learning strategies, ensemble approaches, and emerging paradigms such as federated learning and explainability. Second, we carry out a large benchmarking study on three public datasets: a Kaggle Alzheimer’s disease dataset with four cognitive-impairment stages, a four-class brain tumor dataset, and a seventeen-class brain tumor dataset. We test different model architectures and ensemble methods, and our results provide a clear experimental baseline for future work. For future research, we point out several important open challenges: improving model interpretability, increasing computational efficiency, reducing unfairness and bias, and integrating these tools into real clinical practice. We also suggest several research directions, including self-supervised learning on large neuroimaging datasets, domain-adaptive pretraining, lightweight ensemble distillation, and improved methods for uncertainty estimation. By combining an organized review of recent studies with a systematic benchmarking study, this paper is intended to be both a useful reference for researchers and a guide for advancing clinical AI in neuroimaging. 

Rail buckling remains a major safety concern in modern railway infrastructure, and the effective monitoring of rail temperature is essential for preventing thermal stress failures. However, the direct measurement of rail temperature is often costly, difficult to install in the railway field, and sensitive to various environmental disturbances. This paper presents a new AI-assisted soft sensing approach for indirect rail temperature estimation using a passive thermal sensor consisting of a steel bar with an embedded temperature probe. The approach evaluates seven raw sensor features alongside three engineered features that are designed to capture thermal dynamics and seasonal patterns. To this end, four machine learning models are trained to map the dynamic thermal response of the passive sensor to the underlying rail temperature. Field experiments conducted over a period of five months show that the proposed method achieves robust performance with a mean absolute error of 0.55°C, while reducing both system complexity and installation effort. The results demonstrate that combining passive thermal sensing with data-driven modeling provides a feasible and scalable solution for intelligent rail maintenance and safety monitoring. The approach contributes to AI-assisted instrumentation and measurement by offering a robust soft sensor design that supports improved decision-making for the prevention of rail buckling. 

In this work we introduce AeCSAD, a student–teacher framework designed to detect two types of anomalies in images of industrial components: structural anomalies (physical defects) and logical anomalies (incorrect relationships between components). Unlike prior methods, AeCSAD extends the component segmentation–based logical anomaly detection scheme (c. 2024) with self-attention mechanisms, enabling more effective relational modeling. We demonstrate consistent improvements on the MVTec LOCO Anomaly Detection benchmark. Specifically, AeCSAD employs a global student network with self-attention for reasoning across distant components, complemented by a local student network for fine-grained analysis. Additionally, a patch histogram module measures the frequency distribution of components, allowing the system to detect irregularities in their occurrence. During inference, anomaly scores from the histogram module and the fused local–global networks are combined to produce the final anomaly score. Experiments show that AeCSAD achieves superior average AUROC performance on both structural and logical anomaly detection tasks compared to prior approaches.

The importance of preventing failures in rotating machines has led to extensive research on diagnosis and prognosis methods based on vibration data. To achieve a high generalizability for these solutions, they need to handle high discrepancies between the data used for developing the method and the data from the target machine, a lack of historical events from the target machine and multi-component fault scenarios. Currently, state-of-the-art research has successfully found solutions targeting the first two challenges. However, these almost exclusively focus on single-component fault scenarios, meaning it is unclear how a method should be constructed considering a multi-component system. Therefore, this study constructs a framework for multi-component fault scenarios called Rotating Machinery Prognostic Framework (RoMaP) that leverages the advancements in anomaly detection, fault diagnosis and RUL prediction based on vibration data. To evaluate the potential of RoMaP, an instance based on state-of-the-art research was implemented and compared to other methods based on alternative frameworks. The results show that the instance of RoMaP had the best performance across many scenarios expected in an industrial environment, suggesting that it is a suitable approach for monitoring the health of rotating machines. 

Accurate distance estimation is crucial for applications such as robotics and autonomous systems, where reliable measurements are needed for navigation and interaction with the environment. ArUco markers are a robust solution for distance estimation, offering precise measurements based on their geometric properties. In this paper, we, first, systematically investigate the impact of various contextual factors in an indoor environment, such as angle of observation and illumination, on the systematic error and uncertainty of distance measurements. Our experiments show that illumination (exposure) significantly influence the performance of distance estimation systems, introducing notable systematic errors in real-world settings. Second, we propose a machine learning-based approach, using a neural network, to address the challenge posed by these factors and improve the systematic error of distance estimation. Our results in dynamic outdoor environment, demonstrate the effectiveness of this approach, significantly enhancing the systematic error of distance under dynamic environmental conditions. 

Automated advertisement detection in newspapers is a challenging task due to the diversity in print layouts, formats, and design styles. This task has critical applications in media monitoring, content analysis, and advertising analytics. To address these challenges, we introduce AdVision, a deep-learning-based solution that treats advertisements as unique visual objects. We provide a comparative study of various detection architectures, including one-stage, two-stage, and transformer-based detectors, to identify the most effective approach for detecting advertisements. Our results are validated through extensive experiments conducted under different conditions and metrics. Newspapers from four different countries - Denmark, Norway, Sweden, and the UK - were selected to demonstrate the variety of languages and print formats. Additionally, we conduct a cross-analysis to show how training on one language can generalize to another. To enhance the explainability of our results, we employ GradCAM++ (Chattopadhay et al., 2018) heatmaps. Our experiments demonstrate that the YOLOv8 model achieves superior performance, balancing high precision and recall with minimal inference latency, making it particularly suitable for high-throughput advertisement detection.