Marcella Astrid

Muhammad Zaigham Zaheer, Arif Mahmood, Marcella Astrid et al.

Formulating learning systems for the detection of real-world anomalous events using only video-level labels is a challenging task mainly due to the presence of noisy labels as well as the rare occurrence of anomalous events in the training data. We propose a weakly supervised anomaly detection system which has multiple contributions including a random batch selection mechanism to reduce inter-batch correlation and a normalcy suppression block which learns to minimize anomaly scores over normal regions of a video by utilizing the overall information available in a training batch. In addition, a clustering loss block is proposed to mitigate the label noise and to improve the representation learning for the anomalous and normal regions. This block encourages the backbone network to produce two distinct feature clusters representing normal and anomalous events. Extensive analysis of the proposed approach is provided using three popular anomaly detection datasets including UCF-Crime, ShanghaiTech, and UCSD Ped2. The experiments demonstrate a superior anomaly detection capability of our approach.

Marcella Astrid, Muhammad Zaigham Zaheer, Seung-Ik Lee

Due to the rarity of anomalous events, video anomaly detection is typically approached as one-class classification (OCC) problem. Typically in OCC, an autoencoder (AE) is trained to reconstruct the normal only training data with the expectation that, in test time, it can poorly reconstruct the anomalous data. However, previous studies have shown that, even trained with only normal data, AEs can often reconstruct anomalous data as well, resulting in a decreased performance. To mitigate this problem, we propose to limit the anomaly reconstruction capability of AEs by incorporating pseudo anomalies during the training of an AE. Extensive experiments using five types of pseudo anomalies show the robustness of our training mechanism towards any kind of pseudo anomaly. Moreover, we demonstrate the effectiveness of our proposed pseudo anomaly based training approach against several existing state-ofthe-art (SOTA) methods on three benchmark video anomaly datasets, outperforming all the other reconstruction-based approaches in two datasets and showing the second best performance in the other dataset.

Muhammad Zaigham Zaheer, Jin Ha Lee, Arif Mahmood et al.

Recently, anomaly scores have been formulated using reconstruction loss of the adversarially learned generators and/or classification loss of discriminators. Unavailability of anomaly examples in the training data makes optimization of such networks challenging. Attributed to the adversarial training, performance of such models fluctuates drastically with each training step, making it difficult to halt the training at an optimal point. In the current study, we propose a robust anomaly detection framework that overcomes such instability by transforming the fundamental role of the discriminator from identifying real vs. fake data to distinguishing good vs. bad quality reconstructions. For this purpose, we propose a method that utilizes the current state as well as an old state of the same generator to create good and bad quality reconstruction examples. The discriminator is trained on these examples to detect the subtle distortions that are often present in the reconstructions of anomalous data. In addition, we propose an efficient generic criterion to stop the training of our model, ensuring elevated performance. Extensive experiments performed on six datasets across multiple domains including image and video based anomaly detection, medical diagnosis, and network security, have demonstrated excellent performance of our approach.

Marcella Astrid, Enjie Ghorbel, Djamila Aouada

In this paper, we propose an enhanced audio-visual deep detection method. Recent methods in audio-visual deepfake detection mostly assess the synchronization between audio and visual features. Although they have shown promising results, they are based on the maximization/minimization of isolated feature distances without considering feature statistics. Moreover, they rely on cumbersome deep learning architectures and are heavily dependent on empirically fixed hyperparameters. Herein, to overcome these limitations, we propose: (1) a statistical feature loss to enhance the discrimination capability of the model, instead of relying solely on feature distances; (2) using the waveform for describing the audio as a replacement of frequency-based representations; (3) a post-processing normalization of the fakeness score; (4) the use of shallower network for reducing the computational complexity. Experiments on the DFDC and FakeAVCeleb datasets demonstrate the relevance of the proposed method.

Marcella Astrid, Enjie Ghorbel, Djamila Aouada

Existing methods on audio-visual deepfake detection mainly focus on high-level features for modeling inconsistencies between audio and visual data. As a result, these approaches usually overlook finer audio-visual artifacts, which are inherent to deepfakes. Herein, we propose the introduction of fine-grained mechanisms for detecting subtle artifacts in both spatial and temporal domains. First, we introduce a local audio-visual model capable of capturing small spatial regions that are prone to inconsistencies with audio. For that purpose, a fine-grained mechanism based on a spatially-local distance coupled with an attention module is adopted. Second, we introduce a temporally-local pseudo-fake augmentation to include samples incorporating subtle temporal inconsistencies in our training set. Experiments on the DFDC and the FakeAVCeleb datasets demonstrate the superiority of the proposed method in terms of generalization as compared to the state-of-the-art under both in-dataset and cross-dataset settings.

Marcella Astrid, Enjie Ghorbel, Djamila Aouada

The availability of highly convincing audio deepfake generators highlights the need for designing robust audio deepfake detectors. Existing works often rely solely on real and fake data available in the training set, which may lead to overfitting, thereby reducing the robustness to unseen manipulations. To enhance the generalization capabilities of audio deepfake detectors, we propose a novel augmentation method for generating audio pseudo-fakes targeting the decision boundary of the model. Inspired by adversarial attacks, we perturb original real data to synthesize pseudo-fakes with ambiguous prediction probabilities. Comprehensive experiments on two well-known architectures demonstrate that the proposed augmentation contributes to improving the generalization capabilities of these architectures.

Dat Nguyen, Marcella Astrid, Enjie Ghorbel et al.

Recently, Vision Transformers (ViTs) have achieved unprecedented effectiveness in the general domain of image classification. Nonetheless, these models remain underexplored in the field of deepfake detection, given their lower performance as compared to Convolution Neural Networks (CNNs) in that specific context. In this paper, we start by investigating why plain ViT architectures exhibit a suboptimal performance when dealing with the detection of facial forgeries. Our analysis reveals that, as compared to CNNs, ViT struggles to model localized forgery artifacts that typically characterize deepfakes. Based on this observation, we propose a deepfake detection framework called FakeFormer, which extends ViTs to enforce the extraction of subtle inconsistency-prone information. For that purpose, an explicit attention learning guided by artifact-vulnerable patches and tailored to ViTs is introduced. Extensive experiments are conducted on diverse well-known datasets, including FF++, Celeb-DF, WildDeepfake, DFD, DFDCP, and DFDC. The results show that FakeFormer outperforms the state-of-the-art in terms of generalization and computational cost, without the need for large-scale training datasets. The code is available at \url{https://github.com/10Ring/FakeFormer}.

Dat Nguyen, Enjie Ghorbel, Anis Kacem et al.

In this paper, we propose Localized Artifact Attention X (LAA-X), a novel deepfake detection framework that is both robust to high-quality forgeries and capable of generalizing to unseen manipulations. Existing approaches typically rely on binary classifiers coupled with implicit attention mechanisms, which often fail to generalize beyond known manipulations. In contrast, LAA-X introduces an explicit attention strategy based on a multi-task learning framework combined with blending-based data synthesis. Auxiliary tasks are designed to guide the model toward localized, artifact-prone (i.e., vulnerable) regions. The proposed framework is compatible with both CNN and transformer backbones, resulting in two different versions, namely, LAA-Net and LAA-Former, respectively. Despite being trained only on real and pseudo-fake samples, LAA-X competes with state-of-the-art methods across multiple benchmarks. Code and pre-trained weights for LAA-Net\footnote{https://github.com/10Ring/LAA-Net} and LAA-Former\footnote{https://github.com/10Ring/LAA-Former} are publicly available.

Dat Nguyen, Marcella Astrid, Anis Kacem et al.

Detecting deepfake videos is highly challenging given the complexity of characterizing spatio-temporal artifacts. Most existing methods rely on binary classifiers trained using real and fake image sequences, therefore hindering their generalization capabilities to unseen generation methods. Moreover, with the constant progress in generative Artificial Intelligence (AI), deepfake artifacts are becoming imperceptible at both the spatial and the temporal levels, making them extremely difficult to capture. To address these issues, we propose a fine-grained deepfake video detection approach called FakeSTormer that enforces the modeling of subtle spatio-temporal inconsistencies while avoiding overfitting. Specifically, we introduce a multi-task learning framework that incorporates two auxiliary branches for explicitly attending artifact-prone spatial and temporal regions. Additionally, we propose a video-level data synthesis strategy that generates pseudo-fake videos with subtle spatio-temporal artifacts, providing high-quality samples and hand-free annotations for our additional branches. Extensive experiments on several challenging benchmarks demonstrate the superiority of our approach compared to recent state-of-the-art methods. The code is available at https://github.com/10Ring/FakeSTormer.

Dat Nguyen, Nesryne Mejri, Inder Pal Singh et al.

This paper introduces a novel approach for high-quality deepfake detection called Localized Artifact Attention Network (LAA-Net). Existing methods for high-quality deepfake detection are mainly based on a supervised binary classifier coupled with an implicit attention mechanism. As a result, they do not generalize well to unseen manipulations. To handle this issue, two main contributions are made. First, an explicit attention mechanism within a multi-task learning framework is proposed. By combining heatmap-based and self-consistency attention strategies, LAA-Net is forced to focus on a few small artifact-prone vulnerable regions. Second, an Enhanced Feature Pyramid Network (E-FPN) is proposed as a simple and effective mechanism for spreading discriminative low-level features into the final feature output, with the advantage of limiting redundancy. Experiments performed on several benchmarks show the superiority of our approach in terms of Area Under the Curve (AUC) and Average Precision (AP). The code is available at https://github.com/10Ring/LAA-Net.

Marcella Astrid, Enjie Ghorbel, Djamila Aouada

This paper proposes an audio-visual deepfake detection approach that aims to capture fine-grained temporal inconsistencies between audio and visual modalities. To achieve this, both architectural and data synthesis strategies are introduced. From an architectural perspective, a temporal distance map, coupled with an attention mechanism, is designed to capture these inconsistencies while minimizing the impact of irrelevant temporal subsequences. Moreover, we explore novel pseudo-fake generation techniques to synthesize local inconsistencies. Our approach is evaluated against state-of-the-art methods using the DFDC and FakeAVCeleb datasets, demonstrating its effectiveness in detecting audio-visual deepfakes.

Marcella Astrid, Muhammad Zaigham Zaheer, Seung-Ik Lee

In order to devise an anomaly detection model using only normal training data, an autoencoder (AE) is typically trained to reconstruct the data. As a result, the AE can extract normal representations in its latent space. During test time, since AE is not trained using real anomalies, it is expected to poorly reconstruct the anomalous data. However, several researchers have observed that it is not the case. In this work, we propose to limit the reconstruction capability of AE by introducing a novel latent constriction loss, which is added to the existing reconstruction loss. By using our method, no extra computational cost is added to the AE during test time. Evaluations using three video anomaly detection benchmark datasets, i.e., Ped2, Avenue, and ShanghaiTech, demonstrate the effectiveness of our method in limiting the reconstruction capability of AE, which leads to a better anomaly detection model.

Marcella Astrid, Abdelrahman Shabayek, Djamila Aouada

Batteries are essential for various applications, including electric vehicles and renewable energy storage, making safety and efficiency critical concerns. Anomaly detection in battery thermal images helps identify failures early, but traditional deep learning methods require extensive labeled data, which is difficult to obtain, especially for anomalies due to safety risks and high data collection costs. To overcome this, we explore zero-shot anomaly detection using Visual Question Answering (VQA) models, which leverage pretrained knowledge and textbased prompts to generalize across vision tasks. By incorporating prior knowledge of normal battery thermal behavior, we design prompts to detect anomalies without battery-specific training data. We evaluate three VQA models (ChatGPT-4o, LLaVa-13b, and BLIP-2) analyzing their robustness to prompt variations, repeated trials, and qualitative outputs. Despite the lack of finetuning on battery data, our approach demonstrates competitive performance compared to state-of-the-art models that are trained with the battery data. Our findings highlight the potential of VQA-based zero-shot learning for battery anomaly detection and suggest future directions for improving its effectiveness.

Marcella Astrid, Muhammad Zaigham Zaheer, Djamila Aouada et al.

Due to the rare occurrence of anomalous events, a typical approach to anomaly detection is to train an autoencoder (AE) with normal data only so that it learns the patterns or representations of the normal training data. At test time, the trained AE is expected to well reconstruct normal but to poorly reconstruct anomalous data. However, contrary to the expectation, anomalous data is often well reconstructed as well. In order to further separate the reconstruction quality between normal and anomalous data, we propose creating pseudo anomalies from learned adaptive noise by exploiting the aforementioned weakness of AE, i.e., reconstructing anomalies too well. The generated noise is added to the normal data to create pseudo anomalies. Extensive experiments on Ped2, Avenue, ShanghaiTech, CIFAR-10, and KDDCUP datasets demonstrate the effectiveness and generic applicability of our approach in improving the discriminative capability of AEs for anomaly detection.

Marcella Astrid, Muhammad Zaigham Zaheer, Seung-Ik Lee

Due to the limited availability of anomaly examples, video anomaly detection is often seen as one-class classification (OCC) problem. A popular way to tackle this problem is by utilizing an autoencoder (AE) trained only on normal data. At test time, the AE is then expected to reconstruct the normal input well while reconstructing the anomalies poorly. However, several studies show that, even with normal data only training, AEs can often start reconstructing anomalies as well which depletes their anomaly detection performance. To mitigate this, we propose a temporal pseudo anomaly synthesizer that generates fake-anomalies using only normal data. An AE is then trained to maximize the reconstruction loss on pseudo anomalies while minimizing this loss on normal data. This way, the AE is encouraged to produce distinguishable reconstructions for normal and anomalous frames. Extensive experiments and analysis on three challenging video anomaly datasets demonstrate the effectiveness of our approach to improve the basic AEs in achieving superiority against several existing state-of-the-art models.

Marcella Astrid, Muhammad Zaigham Zaheer, Jae-Yeong Lee et al.

Video anomaly detection is often seen as one-class classification (OCC) problem due to the limited availability of anomaly examples. Typically, to tackle this problem, an autoencoder (AE) is trained to reconstruct the input with training set consisting only of normal data. At test time, the AE is then expected to well reconstruct the normal data while poorly reconstructing the anomalous data. However, several studies have shown that, even with only normal data training, AEs can often start reconstructing anomalies as well which depletes the anomaly detection performance. To mitigate this problem, we propose a novel methodology to train AEs with the objective of reconstructing only normal data, regardless of the input (i.e., normal or abnormal). Since no real anomalies are available in the OCC settings, the training is assisted by pseudo anomalies that are generated by manipulating normal data to simulate the out-of-normal-data distribution. We additionally propose two ways to generate pseudo anomalies: patch and skip frame based. Extensive experiments on three challenging video anomaly datasets demonstrate the effectiveness of our method in improving conventional AEs, achieving state-of-the-art performance.

Jin-Ha Lee, Marcella Astrid, Muhammad Zaigham Zaheer et al.

With the increase in the learning capability of deep convolution-based architectures, various applications of such models have been proposed over time. In the field of anomaly detection, improvements in deep learning opened new prospects of exploration for the researchers whom tried to automate the labor-intensive features of data collection. First, in terms of data collection, it is impossible to anticipate all the anomalies that might exist in a given environment. Second, assuming we limit the possibilities of anomalies, it will still be hard to record all these scenarios for the sake of training a model. Third, even if we manage to record a significant amount of abnormal data, it's laborious to annotate this data on pixel or even frame level. Various approaches address the problem by proposing one-class classification using generative models trained on only normal data. In such methods, only the normal data is used, which is abundantly available and doesn't require significant human input. However, these are trained with only normal data and at the test time, given abnormal data as input, may often generate normal-looking output. This happens due to the hallucination characteristic of generative models. Next, these systems are designed to not use abnormal examples during the training. In this paper, we propose anomaly detection with negative learning (ADNL), which employs the negative learning concept for the enhancement of anomaly detection by utilizing a very small number of labeled anomaly data as compared with the normal data during training. The idea is to limit the reconstruction capability of a generative model using the given a small amount of anomaly examples. This way, the network not only learns to reconstruct normal data but also encloses the normal distribution far from the possible distribution of anomalies.

Muhammad Zaigham Zaheer, Jin-ha Lee, Marcella Astrid et al.

Learning to detect real-world anomalous events using video-level annotations is a difficult task mainly because of the noise present in labels. An anomalous labelled video may actually contain anomaly only in a short duration while the rest of the video can be normal. In the current work, we formulate a weakly supervised anomaly detection method that is trained using only video-level labels. To this end, we propose to utilize binary clustering which helps in mitigating the noise present in the labels of anomalous videos. Our formulation encourages both the main network and the clustering to complement each other in achieving the goal of weakly supervised training. The proposed method yields 78.27% and 84.16% frame-level AUC on UCF-crime and ShanghaiTech datasets respectively, demonstrating its superiority over existing state-of-the-art algorithms.

Muhammad Zaigham Zaheer, Arif Mahmood, Marcella Astrid et al.

Learning to detect real-world anomalous events through video-level labels is a challenging task due to the rare occurrence of anomalies as well as noise in the labels. In this work, we propose a weakly supervised anomaly detection method which has manifold contributions including1) a random batch based training procedure to reduce inter-batch correlation, 2) a normalcy suppression mechanism to minimize anomaly scores of the normal regions of a video by taking into account the overall information available in one training batch, and 3) a clustering distance based loss to contribute towards mitigating the label noise and to produce better anomaly representations by encouraging our model to generate distinct normal and anomalous clusters. The proposed method obtains83.03% and 89.67% frame-level AUC performance on the UCF Crime and ShanghaiTech datasets respectively, demonstrating its superiority over the existing state-of-the-art algorithms.

Muhammad Zaigham Zaheer, Jin-ha Lee, Marcella Astrid et al.

A popular method for anomaly detection is to use the generator of an adversarial network to formulate anomaly scores over reconstruction loss of input. Due to the rare occurrence of anomalies, optimizing such networks can be a cumbersome task. Another possible approach is to use both generator and discriminator for anomaly detection. However, attributed to the involvement of adversarial training, this model is often unstable in a way that the performance fluctuates drastically with each training step. In this study, we propose a framework that effectively generates stable results across a wide range of training steps and allows us to use both the generator and the discriminator of an adversarial model for efficient and robust anomaly detection. Our approach transforms the fundamental role of a discriminator from identifying real and fake data to distinguishing between good and bad quality reconstructions. To this end, we prepare training examples for the good quality reconstruction by employing the current generator, whereas poor quality examples are obtained by utilizing an old state of the same generator. This way, the discriminator learns to detect subtle distortions that often appear in reconstructions of the anomaly inputs. Extensive experiments performed on Caltech-256 and MNIST image datasets for novelty detection show superior results. Furthermore, on UCSD Ped2 video dataset for anomaly detection, our model achieves a frame-level AUC of 98.1%, surpassing recent state-of-the-art methods.

Jeong-Seon Lim, Marcella Astrid, Hyun-Jin Yoon et al.

There are many limitations applying object detection algorithm on various environments. Especially detecting small objects is still challenging because they have low resolution and limited information. We propose an object detection method using context for improving accuracy of detecting small objects. The proposed method uses additional features from different layers as context by concatenating multi-scale features. We also propose object detection with attention mechanism which can focus on the object in image, and it can include contextual information from target layer. Experimental results shows that proposed method also has higher accuracy than conventional SSD on detecting small objects. Also, for 300$\times$300 input, we achieved 78.1% Mean Average Precision (mAP) on the PASCAL VOC2007 test set.

Marcella Astrid, Seung-Ik Lee, Beom-Su Seo

Convolutional Neural Networks (CNNs) is one of successful method in many areas such as image classification tasks. However, the amount of memory and computational cost needed for CNNs inference obstructs them to run efficiently in mobile devices because of memory and computational ability limitation. One of the method to compress CNNs is compressing the layers iteratively, i.e. by layer-by-layer compression and fine-tuning, with CP-decomposition in convolutional layers. To compress with CP-decomposition, rank selection is important. In the previous approach rank selection that is based on sensitivity of each layer, the average rank of the network was still arbitrarily selected. Additionally, the rank of all layers were decided before whole process of iterative compression, while the rank of a layer can be changed after fine-tuning. Therefore, this paper proposes selecting rank of each layer using Variational Bayesian Matrix Factorization (VBMF) which is more systematic than arbitrary approach. Furthermore, to consider the change of each layer's rank after fine-tuning of previous iteration, the method is applied just before compressing the target layer, i.e. after fine-tuning of the previous iteration. The results show better accuracy while also having more compression rate in AlexNet's convolutional layers compression.

Marcella Astrid, Seung-Ik Lee

Convolutional Neural Networks (CNNs) has shown a great success in many areas including complex image classification tasks. However, they need a lot of memory and computational cost, which hinders them from running in relatively low-end smart devices such as smart phones. We propose a CNN compression method based on CP-decomposition and Tensor Power Method. We also propose an iterative fine tuning, with which we fine-tune the whole network after decomposing each layer, but before decomposing the next layer. Significant reduction in memory and computation cost is achieved compared to state-of-the-art previous work with no more accuracy loss.