Nicolae-Catalin Ristea
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Nicolae-Catalin Ristea has a total of 28 co-authors. Here are the top co-authors by number of shared publications:
- Radu Tudor Ionescu
University of Bucharest
8 shared publications - Fahad Shahbaz Khan
Mohamed bin Zayed University of Artificial Intelligence, Abu Dhabi, United Arab Emirates
3 shared publications - Anamaria Radoi
National University of Science and Technology Politehnica Bucharest, Bucharest, Romania
2 shared publications - Ando Saabas
Microsoft Corporation, Redmond, WA, USA
2 shared publications - Mubarak Shah
University of Central Florida
2 shared publications - Nicu Sebe
University of Trento, Trento, Italy
2 shared publications
Latest Publications
- Open Access
- 22 August 2026
- Published by Association for Computing Machinery (ACM) in ACM Computing Surveys
- Vol. 58 (15) , 1-35
Deepfake Media Generation and Detection in the Generative AI Era: A Survey and Outlook
We survey deepfake generation and detection techniques, covering all deepfake media types: image, video, audio and multimodal content. We identify various kinds of deepfakes and construct taxonomies of deepfake generation and detection methods, illustrating the important groups of methods. Next, we gather datasets used for deepfake detection and provide updated rankings of the best performing detectors on the most popular datasets. In addition, we develop a novel multimodal benchmark to evaluate deepfake detectors on out-of-distribution content. The results indicate that state-of-the-art detectors fail to generalize to deepfakes generated by unseen generators. Our project page and new benchmark are available at https://github.com/CroitoruAlin/biodeep.
- 22 August 2026
- Published by Association for Computing Machinery (ACM) in ACM Computing Surveys
- Vol. 58 (15) , 1-35
Open AccessDeepfake Media Generation and Detection in the Generative AI Era: A Survey and Outlook
We survey deepfake generation and detection techniques, covering all deepfake media types: image, video, audio and multimodal content. We identify various kinds of deepfakes and construct taxonomies of deepfake generation and detection methods, illustrating the important groups of methods. Next, we gather datasets used for deepfake detection and provide updated rankings of the best performing detectors on the most popular datasets. In addition, we develop a novel multimodal benchmark to evaluate deepfake detectors on out-of-distribution content. The results indicate that state-of-the-art detectors fail to generalize to deepfakes generated by unseen generators. Our project page and new benchmark are available at https://github.com/CroitoruAlin/biodeep.
- 13 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Geoscience and Remote Sensing Letters
- Vol. 23 (1545598X) , 6010905
Angle-Aware Real-Time Transformers for Oriented Fine-Grained Ship Detection in Optical Remote Sensing Images
Transformer-based detectors, such as the DETR series, have gained significant popularity for object detection due to their strong performance across a variety of benchmarks. However, these architectures are typically limited by high computational costs, restricting their deployment in real-time and resource-constrained environments. To address this limitation, Real-Time DETR (RT-DETR) was introduced, offering an effective trade-off between speed and accuracy. Despite these advances, RT-DETR models are constrained to horizontal bounding boxes and lack the ability to detect oriented objects, a critical requirement in many fields, such as remote sensing. To bridge this gap, we propose AA-RTDETRv2, an angle-aware real-time transformer, specifically designed for oriented object detection and classification. To the best of our knowledge, AA-RTDETRv2 is the first end-to-end RT-DETR-based framework capable of real-time, rotation-aware detection and classification without relying on post-processing steps. Our model introduces two key components: (1) an angle-aware multi-scale encoder that enhances orientation-sensitive feature extraction, and (2) an Angle-Aware Uncertainty-Minimal Query Selection strategy that produces high-quality, rotation-informed object queries to improve detection and classification precision. The proposed AA-RTDETRv2 architecture was evaluated on two standard remote sensing benchmarks (i.e., ShipRSImageNet and DOTA) and achieved state-of-the-art performance-speed balance in oriented object detection and identification. Our method outperforms both NMS-based detectors and prior DETR variants while maintaining inference speeds suitable for real-time applications. These results underscore the effectiveness of AA-RTDETRv2 in addressing remote sensing challenges, including dense object distributions and arbitrary object orientations. The code is accessible at https://to-be-added.
- 13 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Geoscience and Remote Sensing Letters
- Vol. 23 (1545598X) , 6010905
Angle-Aware Real-Time Transformers for Oriented Fine-Grained Ship Detection in Optical Remote Sensing Images
Transformer-based detectors, such as the DETR series, have gained significant popularity for object detection due to their strong performance across a variety of benchmarks. However, these architectures are typically limited by high computational costs, restricting their deployment in real-time and resource-constrained environments. To address this limitation, Real-Time DETR (RT-DETR) was introduced, offering an effective trade-off between speed and accuracy. Despite these advances, RT-DETR models are constrained to horizontal bounding boxes and lack the ability to detect oriented objects, a critical requirement in many fields, such as remote sensing. To bridge this gap, we propose AA-RTDETRv2, an angle-aware real-time transformer, specifically designed for oriented object detection and classification. To the best of our knowledge, AA-RTDETRv2 is the first end-to-end RT-DETR-based framework capable of real-time, rotation-aware detection and classification without relying on post-processing steps. Our model introduces two key components: (1) an angle-aware multi-scale encoder that enhances orientation-sensitive feature extraction, and (2) an Angle-Aware Uncertainty-Minimal Query Selection strategy that produces high-quality, rotation-informed object queries to improve detection and classification precision. The proposed AA-RTDETRv2 architecture was evaluated on two standard remote sensing benchmarks (i.e., ShipRSImageNet and DOTA) and achieved state-of-the-art performance-speed balance in oriented object detection and identification. Our method outperforms both NMS-based detectors and prior DETR variants while maintaining inference speeds suitable for real-time applications. These results underscore the effectiveness of AA-RTDETRv2 in addressing remote sensing challenges, including dense object distributions and arbitrary object orientations. The code is accessible at https://to-be-added.
- Top Cited PapersOpen Access
- 7 January 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Open Journal of Signal Processing
- Vol. 6, 238-246
ICASSP 2024 Speech Signal Improvement Challenge
The ICASSP 2024 Speech Signal Improvement Challenge aims to advance research in enhancing speech signal quality within communication systems. The speech signal quality can be assessed using the SIG metric from ITU-T P.835 and still remains a top issue in audio communication and conferencing systems. For example, in the ICASSP 2023 Deep Noise Suppression Challenge, the improvement in the background and overall quality is impressive, while the speech signal enhancement was not statistically significant. To improve the speech signal the following speech impairment areas must be addressed: coloration, discontinuity, loudness, reverberation, and noise. To this end, we organized ICASSP 2024 Speech Signal Improvement Challenge, which marks the second signal-focused challenge, built upon the success of the previous ICASSP 2023 Speech Signal Improvement Challenge. A training and test set was provided for the challenge, and the winners were determined using an extended crowdsourced implementation of ITU-T P.804's listening phase and the word accuracy (WAcc) rate. The results show that significant improvement was made across all measured dimensions of speech quality.
- 7 January 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Open Journal of Signal Processing
- Vol. 6, 238-246
Top Cited PapersOpen AccessICASSP 2024 Speech Signal Improvement Challenge
The ICASSP 2024 Speech Signal Improvement Challenge aims to advance research in enhancing speech signal quality within communication systems. The speech signal quality can be assessed using the SIG metric from ITU-T P.835 and still remains a top issue in audio communication and conferencing systems. For example, in the ICASSP 2023 Deep Noise Suppression Challenge, the improvement in the background and overall quality is impressive, while the speech signal enhancement was not statistically significant. To improve the speech signal the following speech impairment areas must be addressed: coloration, discontinuity, loudness, reverberation, and noise. To this end, we organized ICASSP 2024 Speech Signal Improvement Challenge, which marks the second signal-focused challenge, built upon the success of the previous ICASSP 2023 Speech Signal Improvement Challenge. A training and test set was provided for the challenge, and the winners were determined using an extended crowdsourced implementation of ITU-T P.804's listening phase and the word accuracy (WAcc) rate. The results show that significant improvement was made across all measured dimensions of speech quality.
- Open Access
- 13 March 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Open Journal of Signal Processing
- Vol. 5, 675-685
ICASSP 2023 Acoustic Echo Cancellation Challenge
The ICASSP 2023 Acoustic Echo Cancellation Challenge is intended to stimulate research in acoustic echo cancellation (AEC), which is an important area of speech enhancement and is still a top issue in audio communication. This is the fourth AEC challenge and it is enhanced by adding a second track for personalized acoustic echo cancellation, reducing the algorithmic + buffering latency to 20 ms, as well as including a full-band version of AECMOS [1]. We open source two large datasets to train AEC models under both single talk and double talk scenarios. These datasets consist of recordings from more than 10,000 real audio devices and human speakers in real environments, as well as a synthetic dataset. We open source an online subjective test framework and provide an objective metric for researchers to quickly test their results. The winners of this challenge were selected based on the average mean opinion score (MOS) achieved across all scenarios and the word accuracy (WAcc) rate.
- 13 March 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Open Journal of Signal Processing
- Vol. 5, 675-685
Open AccessICASSP 2023 Acoustic Echo Cancellation Challenge
The ICASSP 2023 Acoustic Echo Cancellation Challenge is intended to stimulate research in acoustic echo cancellation (AEC), which is an important area of speech enhancement and is still a top issue in audio communication. This is the fourth AEC challenge and it is enhanced by adding a second track for personalized acoustic echo cancellation, reducing the algorithmic + buffering latency to 20 ms, as well as including a full-band version of AECMOS [1]. We open source two large datasets to train AEC models under both single talk and double talk scenarios. These datasets consist of recordings from more than 10,000 real audio devices and human speakers in real environments, as well as a synthetic dataset. We open source an online subjective test framework and provide an objective metric for researchers to quickly test their results. The winners of this challenge were selected based on the average mean opinion score (MOS) achieved across all scenarios and the word accuracy (WAcc) rate.
- Open Access
- 26 October 2023
- Published by Association for Computing Machinery (ACM)
- p. 9467-9471
Cascaded Cross-Modal Transformer for Request and Complaint Detection
- 26 October 2023
- Published by Association for Computing Machinery (ACM)
- p. 9467-9471
Open AccessCascaded Cross-Modal Transformer for Request and Complaint Detection
- 6 October 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 46 (1) , 525-542
Self-Supervised Masked Convolutional Transformer Block for Anomaly Detection
Anomaly detection has recently gained increasing attention in the field of computer vision, likely due to its broad set of applications ranging from product fault detection on industrial production lines and impending event detection in video surveillance to finding lesions in medical scans. Regardless of the domain, anomaly detection is typically framed as a one-class classification task, where the learning is conducted on normal examples only. An entire family of successful anomaly detection methods is based on learning to reconstruct masked normal inputs (e.g. patches, future frames, etc.) and exerting the magnitude of the reconstruction error as an indicator for the abnormality level. Unlike other reconstruction-based methods, we present a novel self-supervised masked convolutional transformer block (SSMCTB) that comprises the reconstruction-based functionality at a core architectural level. The proposed self-supervised block is extremely flexible, enabling information masking at any layer of a neural network and being compatible with a wide range of neural architectures. In this work, we extend our previous self-supervised predictive convolutional attentive block (SSPCAB) with a 3D masked convolutional layer, a transformer for channel-wise attention, as well as a novel self-supervised objective based on Huber loss. Furthermore, we show that our block is applicable to a wider variety of tasks, adding anomaly detection in medical images and thermal videos to the previously considered tasks based on RGB images and surveillance videos. We exhibit the generality and flexibility of SSMCTB by integrating it into multiple state-of-the-art neural models for anomaly detection, bringing forth empirical results that confirm considerable performance improvements on five benchmarks: MVTec AD, BRATS, Avenue, ShanghaiTech, and Thermal Rare Event.
- 6 October 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 46 (1) , 525-542
Self-Supervised Masked Convolutional Transformer Block for Anomaly Detection
Anomaly detection has recently gained increasing attention in the field of computer vision, likely due to its broad set of applications ranging from product fault detection on industrial production lines and impending event detection in video surveillance to finding lesions in medical scans. Regardless of the domain, anomaly detection is typically framed as a one-class classification task, where the learning is conducted on normal examples only. An entire family of successful anomaly detection methods is based on learning to reconstruct masked normal inputs (e.g. patches, future frames, etc.) and exerting the magnitude of the reconstruction error as an indicator for the abnormality level. Unlike other reconstruction-based methods, we present a novel self-supervised masked convolutional transformer block (SSMCTB) that comprises the reconstruction-based functionality at a core architectural level. The proposed self-supervised block is extremely flexible, enabling information masking at any layer of a neural network and being compatible with a wide range of neural architectures. In this work, we extend our previous self-supervised predictive convolutional attentive block (SSPCAB) with a 3D masked convolutional layer, a transformer for channel-wise attention, as well as a novel self-supervised objective based on Huber loss. Furthermore, we show that our block is applicable to a wider variety of tasks, adding anomaly detection in medical images and thermal videos to the previously considered tasks based on RGB images and surveillance videos. We exhibit the generality and flexibility of SSMCTB by integrating it into multiple state-of-the-art neural models for anomaly detection, bringing forth empirical results that confirm considerable performance improvements on five benchmarks: MVTec AD, BRATS, Avenue, ShanghaiTech, and Thermal Rare Event.
- 16 August 2023
- Published by Springer Nature in Applied Intelligence
- Vol. 53 (21) , 25984-26007
Nonlinear neurons with human-like apical dendrite activations
- 16 August 2023
- Published by Springer Nature in Applied Intelligence
- Vol. 53 (21) , 25984-26007
Nonlinear neurons with human-like apical dendrite activations
- Open Access
Non-linear Neurons with Human-like Apical Dendrite Activations
In order to classify linearly non-separable data, neurons are typically organized into multi-layer neural networks that are equipped with at least one hidden layer. Inspired by some recent discoveries in neuroscience, we propose a new model of artificial neuron along with a novel activation function enabling the learning of nonlinear decision boundaries using a single neuron. We show that a standard neuron followed by our novel apical dendrite activation (ADA) can learn the XOR logical function with 100% accuracy. Furthermore, we conduct experiments on six benchmark data sets from computer vision, signal processing and natural language processing, i.e. MOROCO, UTKFace, CREMA-D, Fashion-MNIST, Tiny ImageNet and ImageNet, showing that the ADA and the leaky ADA functions provide superior results to Rectified Linear Units (ReLU), leaky ReLU, RBF and Swish, for various neural network architectures, e.g. one-hidden-layer or two-hidden-layer multi-layer perceptrons (MLPs) and convolutional neural networks (CNNs) such as LeNet, VGG, ResNet and Character-level CNN. We obtain further performance improvements when we change the standard model of the neuron with our pyramidal neuron with apical dendrite activations (PyNADA). Our code is available at: https://github.com/raduionescu/pynada.
Open AccessNon-linear Neurons with Human-like Apical Dendrite Activations
In order to classify linearly non-separable data, neurons are typically organized into multi-layer neural networks that are equipped with at least one hidden layer. Inspired by some recent discoveries in neuroscience, we propose a new model of artificial neuron along with a novel activation function enabling the learning of nonlinear decision boundaries using a single neuron. We show that a standard neuron followed by our novel apical dendrite activation (ADA) can learn the XOR logical function with 100% accuracy. Furthermore, we conduct experiments on six benchmark data sets from computer vision, signal processing and natural language processing, i.e. MOROCO, UTKFace, CREMA-D, Fashion-MNIST, Tiny ImageNet and ImageNet, showing that the ADA and the leaky ADA functions provide superior results to Rectified Linear Units (ReLU), leaky ReLU, RBF and Swish, for various neural network architectures, e.g. one-hidden-layer or two-hidden-layer multi-layer perceptrons (MLPs) and convolutional neural networks (CNNs) such as LeNet, VGG, ResNet and Character-level CNN. We obtain further performance improvements when we change the standard model of the neuron with our pyramidal neuron with apical dendrite activations (PyNADA). Our code is available at: https://github.com/raduionescu/pynada.
- Open Access
SepTr: Separable Transformer for Audio Spectrogram Processing
Following the successful application of vision transformers in multiple computer vision tasks, these models have drawn the attention of the signal processing community. This is because signals are often represented as spectrograms (e.g. through Discrete Fourier Transform) which can be directly provided as input to vision transformers. However, naively applying transformers to spectrograms is suboptimal. Since the axes represent distinct dimensions, i.e. frequency and time, we argue that a better approach is to separate the attention dedicated to each axis. To this end, we propose the Separable Transformer (SepTr), an architecture that employs two transformer blocks in a sequential manner, the first attending to tokens within the same time interval, and the second attending to tokens within the same frequency bin. We conduct experiments on three benchmark data sets, showing that our separable architecture outperforms conventional vision transformers and other state-of-the-art methods. Unlike standard transformers, SepTr linearly scales the number of trainable parameters with the input size, thus having a lower memory footprint. Our code is available as open source at https://github.com/ristea/septr.
Open AccessSepTr: Separable Transformer for Audio Spectrogram Processing
Following the successful application of vision transformers in multiple computer vision tasks, these models have drawn the attention of the signal processing community. This is because signals are often represented as spectrograms (e.g. through Discrete Fourier Transform) which can be directly provided as input to vision transformers. However, naively applying transformers to spectrograms is suboptimal. Since the axes represent distinct dimensions, i.e. frequency and time, we argue that a better approach is to separate the attention dedicated to each axis. To this end, we propose the Separable Transformer (SepTr), an architecture that employs two transformer blocks in a sequential manner, the first attending to tokens within the same time interval, and the second attending to tokens within the same frequency bin. We conduct experiments on three benchmark data sets, showing that our separable architecture outperforms conventional vision transformers and other state-of-the-art methods. Unlike standard transformers, SepTr linearly scales the number of trainable parameters with the input size, thus having a lower memory footprint. Our code is available as open source at https://github.com/ristea/septr.
- Open Access
Self-paced ensemble learning for speech and audio classification
Combining multiple machine learning models into an ensemble is known to provide superior performance levels compared to the individual components forming the ensemble. This is because models can complement each other in taking better decisions. Instead of just combining the models, we propose a self-paced ensemble learning scheme in which models learn from each other over several iterations. During the self-paced learning process based on pseudo-labeling, in addition to improving the individual models, our ensemble also gains knowledge about the target domain. To demonstrate the generality of our self-paced ensemble learning (SPEL) scheme, we conduct experiments on three audio tasks. Our empirical results indicate that SPEL significantly outperforms the baseline ensemble models. We also show that applying self-paced learning on individual models is less effective, illustrating the idea that models in the ensemble actually learn from each other.
Open AccessSelf-paced ensemble learning for speech and audio classification
Combining multiple machine learning models into an ensemble is known to provide superior performance levels compared to the individual components forming the ensemble. This is because models can complement each other in taking better decisions. Instead of just combining the models, we propose a self-paced ensemble learning scheme in which models learn from each other over several iterations. During the self-paced learning process based on pseudo-labeling, in addition to improving the individual models, our ensemble also gains knowledge about the target domain. To demonstrate the generality of our self-paced ensemble learning (SPEL) scheme, we conduct experiments on three audio tasks. Our empirical results indicate that SPEL significantly outperforms the baseline ensemble models. We also show that applying self-paced learning on individual models is less effective, illustrating the idea that models in the ensemble actually learn from each other.