Teacher Network Research Articles

Uncertainty-based knowledge distillation for Bayesian deep neural network compression

Deep learning models have been widely employed across various fields. In real-world scenarios, especially safety-critical applications, quantifying uncertainty is as crucial as achieving high accuracy. To address this concern, Bayesian deep neural networks (BDNNs) emerged to estimate two different types of uncertainty: Aleatoric and Epistemic. Nevertheless, implementing a BDNN on resource-constrained devices poses challenges due to the substantial computational and storage costs imposed by approximation inference techniques. Thus, efficient compression methods should be utilized. We propose an uncertainty-based knowledge distillation method to compress BDNNs. Knowledge distillation is a model compression technique that involves transferring knowledge from a complex network, known as the teacher network, to a simpler one, referred to as the student network. Our method incorporates uncertainty into knowledge distillation to address situations where inappropriate teacher supervision undermines compression performance. We utilize the Epistemic uncertainty of teacher predictions to tailor supervision for each sample individually to take into account teacher's limited knowledge. Additionally, we adjust the temperature parameter of the distillation process for each sample based on the Aleatoric uncertainty of the teacher predictions, ensuring that the student receives appropriate supervision even in the presence of ambiguous data. As a result, the proposed method enables the Bayesian student network to be trained under both appropriate supervision of the Bayesian teacher network and ground truth labels. We evaluated our method on the CIFAR-10, CIFAR-100, and RAF-DB datasets, demonstrating notable improvements in accuracy over state-of-the-art knowledge distillation-based methods. Furthermore, the robustness of our approach was assessed through testing weakly trained teacher networks and the analysis of blurred and low-resolution data, which have high uncertainty. Experimental results show that the proposed method outperformed existing methods.

The Relationship between Resiliency, Psychological Empowerment, and Teacher Burnout across Different Genders: A Psychological Network Analysis.

Teacher burnout is one of the main reasons leading to decreased teaching performance and occupational mental health issues among teachers, drawing widespread global attention. Previous research has found that both resiliency and psychological empowerment can alleviate teacher burnout, yet there is no study simultaneously examining the relationships between resiliency, PE, and teacher burnout. Furthermore, previous studies have found gender differences in teacher burnout but have not examined the moderating effects of resiliency and psychological empowerment by gender group. Additionally, traditional analytical methods may overlook the compositional connections between these variables. To bridge this gap, we employed psychological network analysis to evaluate the psychological network of teachers with burnout across different genders. Findings indicate: (1) Female teachers exhibit a stronger link between their sense of departmental control and burnout, while male teachers show a stronger connection between solving instructional challenges and burnout. (2) Patience in male teachers' approach to teaching may enhance connections with their environment, and mastering job-relevant skills can boost male teachers' job happiness. (3) Female teachers' "I feel connected to others" demonstrates higher bridge centrality. In comparison, male teachers' "My work is vital to me" shows higher bridge centrality, indicating deeper connections with other symptom clusters. This study reveals the complex interactions among the factors of teacher burnout and investigates how gender differences influence the associations between these factors and burnout, by not only filling gaps in previous research but also offering new perspectives and strategies for understanding and intervening in teacher burnout, especially in the context of gender differences.

LG-AKD: Application of a lightweight GCN model based on adversarial knowledge distillation to skeleton action recognition

Human action recognition, a pivotal topic in computer vision, is a highly complex and challenging task. It requires the analysis of not only spatial dependencies of targets but also temporal changes in these targets. In recent decades, the advancement of deep learning has led to the development of numerous action recognition methods based on deep neural networks. Given that the skeleton points of the human body can be treated as a graph structure, graph neural networks (GNNs) have emerged as an effective tool for modeling such data, garnering significant interest from researchers. This paper aims to address the issue of low test speed caused by over-complicated deep graph convolutional models. To achieve this, we compress the network structure using knowledge distillation from a teacher-student architecture, leading to a compact and lightweight student GNN. To enhance the model’s robustness and generalization capabilities, we introduce a data augmentation mechanism that generates diverse action sequences while maintaining consistent behavior labels, thereby providing a more comprehensive learning basis for the model. The proposed model integrates three distinct knowledge learning paths: teacher networks, original datasets, and derived data. The fusion of knowledge distillation and data augmentation enables lightweight student networks to outperform their teacher networks in terms of both performance and efficiency. Experimental results demonstrate the efficacy of our approach in the context of skeleton-based human action recognition, highlighting its potential to simplify state-of-the-art models while enhancing their performance.

Efficient face anti-spoofing via head-aware transformer based knowledge distillation with 5 MB model parameters

Although face recognition technology has been applied in many scenarios, it still suffers from many types of presentation attacks, so face anti-spoofing (FAS) becomes a hot topic in computer vision. Recently, vision transformer is recognized as the mainstream architecture for FAS, which always relies on auxiliary information, sophisticated tricks and huge model parameters. Considering that face based identity authentication usually takes place on mobile-like devices, therefore how to design an effective and lightweight model is of great significance. Inspired by the powerful global modeling ability of self-attention and the model compression ability of knowledge distillation, a simple yet effective knowledge distillation approach is proposed for FAS under transformer framework. Our primary idea is to leverage the rich knowledge of a teacher network pre-trained on large-scale face data to guide the learning of a lightweight student network. The main contributions of our method are threefold: (1) Feature- and logits-level distillation are combined to transfer the rich knowledge of teacher to student. (2) A head-aware strategy is proposed to deal with the dimension mismatching issue of middle encoder layers between teacher and student networks, in which a novel attention head correlation matrix is introduced. (3) Our method can bridge the performance gap between teacher and student, and the resulting student network is extremely lightweight with only 5 MB parameters. Extensive experiments are conducted on three public face-spoofing datasets, CASIA-FASD, Replay-Attack and OULU-NPU, the results demonstrate that our method can obtain performance on par with or superior to most FAS methods and outperform many knowledge distillation methods. Meanwhile, the distilled student network achieves excellent performance with 17× fewer parameters and 9× faster inference time compared to the teacher network. The code will be publicly available at https://github.com/Maricle-zhangjun/HaTFAS.

Towards accurate diagnosis: exploring knowledge distillation and self-attention in multimodal medical image fusion

ABSTRACT Multimodal medical image fusion aims to aggregate significant information based on the characteristics of medical images from different modalities. Existing research in image fusion faces several major limitations, including a scarcity of paired data, noisy and inconsistent modalities, a lack of contextual relationships, and suboptimal feature extraction and fusion techniques. In response to these challenges, this research proposes a novel adaptive fusion approach. Our knowledge distillation (KD) model extracts informative features from multimodal medical images using various key components. A teacher network is employed to emphasise the suitability and complexity of capturing high-level abstract features. The soft labels are utilised to transfer the knowledge between the teacher network as well as the student network. During student network training, we minimise the divergence between these soft labels. To enhance the adaptive fusion of extracted features from different modalities, we apply a self-attention mechanism. Training this self-attention mechanism minimises the loss function, encouraging attention scores to capture relevant contextual relationships between features. Additionally, a cross-modal consistency module aligns the extracted features to ensure spatial consistency and meaningful fusion. Our adaptive fusion strategy effectively combines features to enhance the diagnostic value and quality of fused images. We employ generator and discriminator architectures for synthesising fused images and distinguishing between real and generated fused images. Comprehensive analysis is conducted on the basis of diverse evaluation measures. Experimental results demonstrate improved fusion outcomes with values of 0.92, 41.58, 7.25, 0.958, 0.759, 0.947, 0.90, 7.05, 0.0726, and 76 s for SSIM, PSNR, FF, VIF, UIQI, FMI, EITF, entropy, RMSE, and execution time, respectively.

NTS-CAM classification model with channel attention mechanism for grading In-Vitro Fertilization (IVF) blastocyst quality

An automated-based intelligence approaches have widely used for quantifying In-Vitro Fertilisation (IVF) blastocyst image features that offer automation in morphology assessment as well as embryo selection to improve embryo implantation. Since the IVF blastocyst co-existed three main features of Zona Pellucida (ZP), Trophectoderm (TE) and Inner Cell Mass (ICM), this has made it crucial to consider the informative regions of all features in image morphology assessment. Although the implementation of Navigator-Teacher-Scrutinizer Network (NTS-net) has been detected most informative regions under the guidance of the Teacher network, there still limitation on calculation of the feature extraction process of different blastocyst features that led to poor classification performance. Therefore, this study proposes a new classification model namely NTS-CAM to improve extracted blastocyst features by assigning weights to channel features in channel attention mechanism (CAM) while extracting informative regions of each blastocyst feature. The benchmarking dataset showed significant performance of classification accuracy for ZP, TE, and ICM features with 80.5 %, 67.4 %, and 76.3 %, and the clinical dataset showed 74.1 %, 71.8 %, and 63.5 %, respectively. In conclusion, the proposed NTS-CAM model to predict grade of IVF blastocyst quality has improved the performance compared to classic NTS model. Furthermore, the improved model can be used for clinical decision making as well as for quality control in IVF procedure.

Distilling Privileged Knowledge for Anomalous Event Detection From Weakly Labeled Videos.

Weakly supervised video anomaly detection (WS-VAD) aims to identify the snippets involving anomalous events in long untrimmed videos, with solely text video-level binary labels. A typical paradigm among the existing text WS-VAD methods is to employ multiple modalities as inputs, e.g., RGB, optical flow, and audio, as they can provide sufficient discriminative clues that are robust to the diverse, complicated real-world scenes. However, such a pipeline has high reliance on the availability of multiple modalities and is computationally expensive and storage demanding in processing long sequences, which limits its use in some applications. To address this dilemma, we propose a privileged knowledge distillation (KD) framework dedicated to the WS-VAD task, which can maintain the benefits of exploiting additional modalities, while avoiding the need for using multimodal data in the inference phase. We argue that the performance of the privileged KD framework mainly depends on two factors: 1) the effectiveness of the multimodal teacher network and 2) the completeness of the useful information transfer. To obtain a reliable teacher network, we propose a text cross-modal interactive learning strategy and an anomaly normal discrimination loss, which target learning task-specific cross-modal features and encourage the separability of anomalous and normal representations, respectively. Furthermore, we design both representation- and text logits-level distillation loss functions, which force the unimodal student network to distill abundant privileged knowledge from the text well-trained multimodal teacher network, in a snippet-to-video fashion. Extensive experimental results on three public benchmarks demonstrate that the proposed privileged KD framework can train a lightweight yet effective detector, for localizing anomaly events under the supervision of video-level annotations.

Efficient skin lesion segmentation with boundary distillation.

Medical image segmentation models are commonly known for their complex structures, which often render them impractical for use on edge computing devices and compromising efficiency in the segmentation process. In light of this, the industry has proposed the adoption of knowledge distillation techniques. Nevertheless, the vast majority of existing knowledge distillation methods are focused on the classification tasks of skin diseases. Specifically, for the segmentation tasks of dermoscopy lesion images, these knowledge distillation methods fail to fully recognize the importance of features in the boundary regions of lesions within medical images, lacking boundary awareness for skin lesions. This paper introduces pioneering medical image knowledge distillation architecture. The aim of this method is to facilitate the efficient transfer of knowledge from existing complex medical image segmentation networks to a more simplified student network. Initially, a masked boundary feature (MBF) distillation module is designed. By applying random masking to the periphery of skin lesions, the MBF distillation module obliges the student network to reproduce the comprehensive features of the teacher network. This process, in turn, augments the representational capabilities of the student network. Building on the MBF distillation module, this paper employs a cascaded combination approach to integrate the MBF distillation module into a multi-head boundary feature (M2BF) distillation module, further strengthening the student network's feature learning ability and enhancing the overall image segmentation performance of the distillation model. This method has been experimentally validated on the public datasets ISIC-2016 and PH2, with results showing significant performance improvements in the student network. Our findings highlight the practical utility of the lightweight network distilled using our approach, particularly in scenarios demanding high operational speed and minimal storage usage. This research offers promising prospects for practical applications in the realm of medical image segmentation.

Cross-Modal Consistency for Single-Modal MR Image Segmentation.

Multi-modal magnetic resonance (MR) image segmentation is an important task in disease diagnosis and treatment, but it is usually difficult to obtain multiple modalities for a single patient in clinical applications. To address these issues, a cross-modal consistency framework is proposed for a single-modal MR image segmentation. To enable single-modal MR image segmentation in the inference stage, a weighted cross-entropy loss and a pixel-level feature consistency loss are proposed to train the target network with the guidance of the teacher network and the auxiliary network. To fuse dual-modal MR images in the training stage, the cross-modal consistency is measured according to Dice similarity entropy loss and Dice similarity contrastive loss, so as to maximize the prediction similarity of the teacher network and the auxiliary network. To reduce the difference in image contrast between different MR images for the same organs, a contrast alignment network is proposed to align input images with different contrasts to reference images with a good contrast. Comprehensive experiments have been performed on a publicly available prostate dataset and an in-house pancreas dataset to verify the effectiveness of the proposed method. Compared to state-of-the-art methods, the proposed method can achieve better segmentation. The proposed image segmentation method can fuse dual-modal MR images in the training stage and only need one-modal MR images in the inference stage. The proposed method can be used in routine clinical occasions when only single-modal MR image with variable contrast is available for a patient.

Adaptive Deep Neural Network for Click-Through Rate estimation

Over the past few years, deep learning networks have been applied in the estimation of the Click-Through Rate (CTR). The main task of the CTR model lies in predicting a user’s positive/negative response to an item. Most of existing CTR models are constituted of two parts: a deep neural network (DNN) and a wide model (e.g., deep cross network). The wide models are generally designed to learn the feature interaction and contain much more parameters than the DNN. As a consequence, it is quite difficult for existing CTR models to work on the resource-limited devices. Blindly removing the wide model will significantly encumber the predictive performance of the algorithm. In this paper, we adopt the knowledge distillation to train a single multi-branch network and assemble all branches as a teacher network. Each branch network is a simple DNN which not only matches the ground-truth label distribution but also aligns the prediction distribution of the teacher network. One of branch networks, termed as Adaptive Deep Neural Network (ADNN), is trained independently and further combined with a wide model to learn feature interactions. Our method does not require pre-training any high-capacity teacher models, which endows our method with higher efficiency compared with existing ones. The experimental results tested on Criteo and Avazu datasets show that the hybrid model outperforms state-of-the-art methods, and the light model ADNN also has a considerable performance accuracy over certain modern complex models, demonstrating the superiority of the methodology.

A framework for the emergence and analysis of language in social learning agents

Neural systems have evolved not only to solve environmental challenges through internal representations but also, under social constraints, to communicate these to conspecifics. In this work, we aim to understand the structure of these internal representations and how they may be optimized to transmit pertinent information from one individual to another. Thus, we build on previous teacher-student communication protocols to analyze the formation of individual and shared abstractions and their impact on task performance. We use reinforcement learning in grid-world mazes where a teacher network passes a message to a student to improve task performance. This framework allows us to relate environmental variables with individual and shared representations. We compress high-dimensional task information within a low-dimensional representational space to mimic natural language features. In coherence with previous results, we find that providing teacher information to the student leads to a higher task completion rate and an ability to generalize tasks it has not seen before. Further, optimizing message content to maximize student reward improves information encoding, suggesting that an accurate representation in the space of messages requires bi-directional input. These results highlight the role of language as a common representation among agents and its implications on generalization capabilities.

Self-Adaptive Teacher-Student framework for colon polyp segmentation from unannotated private data with public annotated datasets.

Colon polyps have become a focal point of research due to their heightened potential to develop into appendiceal cancer, which has the highest mortality rate globally. Although numerous colon polyp segmentation methods have been developed using public polyp datasets, they tend to underperform on private datasets due to inconsistencies in data distribution and the difficulty of fine-tuning without annotations. In this paper, we propose a Self-Adaptive Teacher-Student (SATS) framework to segment colon polyps from unannotated private data by utilizing multiple publicly annotated datasets. The SATS trains multiple teacher networks on public datasets and then generates pseudo-labels on private data to assist in training a student network. To enhance the reliability of the pseudo-labels from the teacher networks, the SATS includes a newly proposed Uncertainty and Distance Fusion (UDFusion) strategy. UDFusion dynamically adjusts the pseudo-label weights based on a novel reconstruction similarity measure, innovatively bridging the gap between private and public data distributions. To ensure accurate identification and segmentation of colon polyps, the SATS also incorporates a Granular Attention Network (GANet) architecture for both teacher and student networks. GANet first identifies polyps roughly from a global perspective by encoding long-range anatomical dependencies and then refines this identification to remove false-positive areas through multi-scale background-foreground attention. The SATS framework was validated using three public datasets and one private dataset, achieving 76.30% on IoU, 86.00% on Recall, and 7.01 pixels on HD. These results outperform the existing five methods, indicating the effectiveness of this approach for colon polyp segmentation.

Lightweight transmission line defect identification method based on OFN network and distillation method

AbstractEdge devices are increasingly utilized for defect detection in power line inspection, necessitating algorithms that optimize model size and accuracy. This study introduces a lightweight detection method using the Optimized Feature Network with MobileOne‐FPN‐NASHead (OFN network) OFN network and distillation technique. The OFN incorporates a lightweight backbone, neural network search, re‐parametrization, and a feature pyramid to create a compact yet effective detection network. To enhance feature learning, a heterogeneous distillation approach is applied, leveraging a modified YOLOv8 as a teacher network. This modification includes an explicit visual centre for improved global and local information extraction, crucial for dense target detection in power line inspections. Additionally, the Minimize the points distance IoU (MPDloU) loss function is used to improve localization accuracy over the the Complete‐Intersection Over Union (CIoU) loss. Experimental results show a 1.1% mean Average Precision (mAP) increase for the enhanced YOLOv8 and a 70.2% mAP for the OFN network with 18.95 GFLOPs and 343 FPS, achieving a commendable balance between model efficiency and detection performance. The research underscores the viability of the OFN for edge computing in power line defect detection, highlighting the potential of innovative algorithmic structures in this application.

Distilling interaction knowledge for semi-supervised egocentric action recognition

Egocentric action recognition, the identification of actions within video content obtained from a first-person perspective, is receiving increasing attention due to the widespread adoption of wearable camera technology. Nonetheless, the task of annotating actions within a video characterized by a cluttered background and the presence of various objects is labor-intensive. In this paper, we consider learning for egocentric action recognition in a semi-supervised manner. Inspired by the fact that videos captured from first-person viewpoint usually contain rich contents about how human hands interact with objects, we thus propose to employ a popular teacher–student framework and distill the interaction knowledge between hand and objects for semi-supervised egocentric action recognition. We refer to the proposed method as Interaction Knowledge Distillation or IKD. Specifically, the teacher network takes hands and action-related objects in the labeled videos as input, and uses graph neural networks to capture their spatial–temporal relations as graph edge features. The student network then takes the detected hands/objects from both labeled and unlabeled videos as input and mimics the teacher network to learn from the interactions to improve model performance. Experiments are performed on two popular egocentric action recognition datasets, Something-Something-V2 and EPIC-KITCHENS-100, which show that our proposed approach consistently outperforms recent state-of-the-art methods in typical semi-supervised settings.

Lightweight intrusion detection model based on CNN and knowledge distillation

The problem of network attacks is a primary focus in the domain of intrusion detection. Models face significant challenges in recognizing intrusion behaviors, particularly when dealing with high-dimensional and sparse datasets. Traditional machine learning methods often struggle with these dimensionality issues. In contrast, deep learning, a crucial technology in intrusion detection, excels at managing high-dimensional data. However, traditional image coding methods do not adequately address data sparsification and often overlook the spatial continuity among features. The Fourier transform is a promising solution for data sparsity issues, as it effectively mitigates the impact by converting data into a different domain. Inspired by the Fourier transform, this paper proposes a lightweight intrusion detection model called TFTKD, based on Convolutional Neural Networks (CNN) and knowledge distillation. The model applies a two-dimensional Fourier transform to convert grayscale images from the time domain to the frequency domain. This transformation enhances the similarity between neighboring pixels, effectively addressing data sparsification. During the training phase, a teacher network, comprising an 8-layer CNN, is pre-trained. In the distillation phase, a one-layer CNN serves as the student network, employing Self-adaptive Temperature Knowledge Distillation to enhance the student's generalization capabilities. This approach results in a compact student network model with a constrained parameter count, demonstrating superior learning efficiency and accuracy compared to state-of-the-art methods. Experimental validation was conducted using four publicly accessible intrusion detection datasets, demonstrating the effectiveness of the proposed method.

Semi-supervised abdominal multi-organ segmentation by object-redrawing.

Multi-organ segmentation is a critical task in medical imaging, with wide-ranging applications in both clinical practice and research. Accurate delineation of organs from high-resolution 3D medical images, such as CT scans, is essential for radiation therapy planning, enhancing treatment outcomes, and minimizing radiation toxicity risks. Additionally, it plays a pivotal role in quantitative image analysis, supporting various medical research studies. Despite its significance, manual segmentation of multiple organs from 3D images is labor-intensive and prone to low reproducibility due to high interoperator variability. Recent advancements in deep learning have led to several automated segmentation methods, yet many rely heavily on labeled data and human anatomy expertise. In this study, our primary objective is to address the limitations of existing semi-supervised learning (SSL) methods for abdominal multi-organ segmentation. We aim to introduce a novel SSL approach that leverages unlabeled data to enhance the performance of deep neural networks in segmenting abdominal organs. Specifically, we propose a method that incorporates a redrawing network into the segmentation process to correct errors and improve accuracy. Our proposed method comprises three interconnected neural networks: a segmentation network for image segmentation, a teacher network for consistency regularization, and a redrawing network for object redrawing. During training, the segmentation network undergoes two rounds of optimization: basic training and readjustment. We adopt the Mean-Teacher model as our baseline SSL approach, utilizing labeled and unlabeled data. However, recognizing significant errors in abdominal multi-organ segmentation using this method alone, we introduce the redrawing network to generate redrawn images based on CT scans, preserving original anatomical information. Our approach is grounded in the generative process hypothesis, encompassing segmentation, drawing, and assembling stages. Correct segmentation is crucial for generating accurate images. In the basic training phase, the segmentation network is trained using both labeled and unlabeled data, incorporating consistency learning to ensure consistent predictions before and after perturbations. The readjustment phase focuses on reducing segmentation errors by optimizing the segmentation network parameters based on the differences between redrawn and original CT images. We evaluated our method using two publicly available datasets: the beyond the cranial vault (BTCV) segmentation dataset (training: 44, validation: 6) and the abdominal multi-organ segmentation (AMOS) challenge 2022 dataset (training:138, validation:16). Our results were compared with state-of-the-art SSL methods, including MT and dual-task consistency (DTC), using the Dice similarity coefficient (DSC) as an accuracy metric. On both datasets, our proposed SSL method consistently outperformed other methods, including supervised learning, achieving superior segmentation performance for various abdominal organs. These findings demonstrate the effectiveness of our approach, even with a limited number of labeled data. Our novel semi-supervised learning approach for abdominal multi-organ segmentation addresses the challenges associated with this task. By integrating a redrawing network and leveraging unlabeled data, we achieve remarkable improvements in accuracy. Our method demonstrates superior performance compared to existing SSL and supervised learning methods. This approach holds great promise in enhancing the precision and efficiency of multi-organ segmentation in medical imaging applications.

Semi-supervised learning for efficient water leakage segmentation in tunnel infrastructure

Water leakage segmentation based on computer vision allows for efficient monitoring and maintenance, ensuring the safety and integrity of tunnel infrastructure. In this study, a semisupervised learning method with self-training via pseudo-labels is proposed to overcome the dependence of conventional computer vision models on high-quality labeled data. Strong data augmentation is first injected into the unlabeled data to establish a robust self-training baseline and differentiate similar predictions between teacher and student networks. To mitigate the accumulation of erroneous pseudo-labels and their potential impact on performance, we chose to retrain using only reliable, unlabeled images. The unlabeled images are prioritized based on their overall stability. By considering image-level contextual information, the selection process offers a more suitable approach for segmentation compared to conventional pixelwise methods. The experimental results highlight the method’s capability to enhance model performance with a limited amount of labeled data, surpassing the effectiveness of other semisupervised methods.

Action research in non-formal education, outreach projects with networks of teachers in Michoacán, Mexico

In this study, we analyze the intention to implement workshops focused on basic concepts of Astronomy by elementary school teachers in their classrooms. An analysis was conducted on a sample, and the variables that influence the intention to implement the workshops by a group of teachers from the state of Michoacán are reviewed. The objective is to create a diagnosis that allows obtaining information to be used in a subsequent action research study, with the aim of generating improvements in the workshop offerings targeted at teachers so that they can replicate them.

MST: Multiscale Flow-Based Student–Teacher Network for Unsupervised Anomaly Detection

Student–teacher networks have shown promise in unsupervised anomaly detection; however, issues such as semantic confusion and abnormal deformations still restrict the detection accuracy. To address these issues, we propose a novel student–teacher network named MST by integrating the multistage pixel-reserving bridge (MPRB) and the spatial compression autoencoder (SCA) to the MMR network. The MPRB enhances inter-level information interaction and local feature extraction, improving the anomaly localization and reducing the false detection area. The SCA bolsters global feature extraction, making the detection boundaries of larger defects clearer. By testing our network across various datasets, our method achieves state-of-the-art (SOTA) performance on AeBAD-S, AeBAD-V, and MPDD datasets, with image-level AUROC scores of 87.5%, 78.5%, and 96.5%, respectively. Furthermore, our method also exhibits competitive performance on the widely utilized MVTec AD dataset.

Dual-branch teacher-student with noise-tolerant learning for domain adaptive nighttime segmentation

While significant progress has been achieved in the field of image semantic segmentation, the majority of research has been primarily concentrated on daytime scenes. Semantic segmentation of nighttime images is equally critical for autonomous driving; however, this task presents greater challenges due to inadequate lighting and difficulties associated with obtaining accurate manual annotations. In this paper, we introduce a novel method called the Dual-Branch Teacher-Student (DBTS) framework for unsupervised nighttime semantic segmentation. Our approach combines domain alignment and knowledge distillation in a mutually reinforcing manner. Firstly, we employ a photometric alignment module to dynamically generate target-like latent images, bridging the appearance gap between the source domain (daytime) and the target domain (nighttime). Secondly, we establish a dual-branch framework, where each branch enhances collaboration between the teacher and student networks. The student network utilizes adversarial learning to align the target domain with another domain (i.e., source or latent domain), while the teacher network generates reliable pseudo-labels by distilling knowledge from the latent domain. Furthermore, recognizing the potential noise present in pseudo-labels, we propose a noise-tolerant learning method to mitigate the risks associated with overreliance on pseudo-labels during domain adaptation. When evaluated on benchmark datasets, the proposed DBTS achieves state-of-the-art performance. Specifically, DBTS, using different backbones, outperforms established baseline models by approximately 25% in mIoU on the Zurich dataset and by over 26% in mIoU on the ACDC dataset, demonstrating the effectiveness of our method in addressing the challenges of domain-adaptive nighttime segmentation.