Generation of images containing multiple humans, performing complex actions, while preserving their facial identities, is a significant challenge. A major factor contributing to this is the lack of a a dedicated benchmark. To address this, we introduce MultiHuman-Testbench, a novel benchmark for rigorously evaluating generative models for multi-human generation. The benchmark comprises 1800 samples, including carefully curated text prompts, describing a range of simple to complex human actions. These prompts are matched with a total of 5,550 unique human face images, sampled uniformly to ensure diversity across age, ethnic background, and gender. Alongside captions, we provide human-selected pose conditioning images which accurately match the prompt. We propose a multi-faceted evaluation suite employing four key metrics to quantify face count, ID similarity, prompt alignment, and action detection. We conduct a thorough evaluation of a diverse set of models, including zero-shot approaches and training-based methods, with and without regional priors. We also propose novel techniques to incorporate image and region isolation using human segmentation and Hungarian matching, significantly improving ID similarity. Our proposed benchmark and key findings provide valuable insights and a standardized tool for advancing research in multi-human image generation.
A great amount of endeavor has recently been devoted to activity detection for massive machine-type communications in cell-free massive MIMO. However, in practice, as the number of antennas at the access points (APs) increases, the Rayleigh distance that separates the near-field and far-field regions also expands, rendering the conventional assumption of far-field propagation alone impractical. To address this challenge, this paper considers a hybrid near-far field activity detection in cell-free massive MIMO, and establishes a covariance-based formulation, which facilitates the development of a distributed algorithm to alleviate the computational burden at the central processing unit (CPU). Specifically, each AP performs local activity detection for the devices and then transmits the detection result to the CPU for further processing. In particular, a novel coordinate descent algorithm based on the Sherman-Morrison-Woodbury update with Taylor expansion is proposed to handle the local detection problem at each AP. Moreover, we theoretically analyze how the hybrid near-far field channels affect the detection performance. Simulation results validate the theoretical analysis and demonstrate the superior performance of the proposed approach compared with existing approaches.
Automatic fluency assessment (AFA) remains challenging, particularly in capturing speech rhythm, pauses, and disfluencies in non-native speakers. We introduce a chunk-based approach integrating self-supervised learning (SSL) models (Wav2Vec2, HuBERT, and WavLM) selected for their complementary strengths in phonetic, prosodic, and noisy speech modeling, with a hierarchical CNN-BiLSTM framework. Speech is segmented into breath-group chunks using Silero voice activity detection (Silero-VAD), enabling fine-grained temporal analysis while mitigating over-segmentation artifacts. SSL embeddings are fused via a learnable weighted mechanism, balancing acoustic and linguistic features, and enriched with chunk-level fluency markers (e.g., speech rate, pause durations, n-gram repetitions). The CNN-BiLSTM captures local and long-term dependencies across chunks. Evaluated on Avalinguo and Speechocean762, our approach improves F1-score by 2.8 and Pearson correlation by 6.2 points over single SSL baselines on Speechocean762, with gains of 4.2 F1-score and 4.0 Pearson points on Avalinguo, surpassing Pyannote.audio-based segmentation baselines. These findings highlight chunk-based multi-SSL fusion for robust fluency evaluation, though future work should explore generalization to dialects with irregular prosody.
Massive connectivity supports the sporadic access of a vast number of devices without requiring prior permission from the base station (BS). In such scenarios, the BS must perform joint activity detection and channel estimation (JADCE) prior to data reception. Message passing algorithms have emerged as a prominent solution for JADCE under a Bayesian inference framework. The existing message passing algorithms, however, typically rely on some hand-crafted and overly simplistic priors of the wireless channel, leading to significant channel estimation errors and reduced activity detection accuracy. In this paper, we focus on the problem of JADCE in a multiple-input multiple-output orthogonal frequency division multiplexing (MIMO-OFDM) grant-free random access network. We propose to incorporate a more accurate channel prior learned by score-based generative models into message passing, so as to push towards the performance limit of JADCE. Specifically, we develop a novel turbo message passing (TMP) framework that models the entire channel matrix as a super node, rather than factorizing it element-wise. This design enables the seamless integration of score-based generative models as a minimum mean-squared error (MMSE) denoiser. The variance of the denoiser, which is essential in message passing, can also be learned through score-based generative models. Our approach, termed score-based TMP for JADCE (STMP-JADCE), takes full advantages of the powerful generative prior and, meanwhile, benefits from the fast convergence speed of message passing. Numerical simulations show that STMP-JADCE drastically enhances the activity detection and channel estimation performance compared to the state-of-the-art baseline algorithms.
In speaker diarization, traditional clustering-based methods remain widely used in real-world applications. However, these methods struggle with the complex distribution of speaker embeddings and overlapping speech segments. To address these limitations, we propose an Overlapping Community Detection method based on Graph Attention networks and the Label Propagation Algorithm (OCDGALP). The proposed framework comprises two key components: (1) a graph attention network that refines speaker embeddings and node connections by aggregating information from neighboring nodes, and (2) a label propagation algorithm that assigns multiple community labels to each node, enabling simultaneous clustering and overlapping community detection. Experimental results show that the proposed method significantly reduces the Diarization Error Rate (DER), achieving a state-of-the-art 15.94% DER on the DIHARD-III dataset without oracle Voice Activity Detection (VAD), and an impressive 11.07% with oracle VAD.
Voice Activity Detection (VAD) plays a key role in speech processing, often utilizing hand-crafted or neural features. This study examines the effectiveness of Mel-Frequency Cepstral Coefficients (MFCCs) and pre-trained model (PTM) features, including wav2vec 2.0, HuBERT, WavLM, UniSpeech, MMS, and Whisper. We propose FusionVAD, a unified framework that combines both feature types using three fusion strategies: concatenation, addition, and cross-attention (CA). Experimental results reveal that simple fusion techniques, particularly addition, outperform CA in both accuracy and efficiency. Fusion-based models consistently surpass single-feature models, highlighting the complementary nature of MFCCs and PTM features. Notably, our best-performing fusion model exceeds the state-of-the-art Pyannote across multiple datasets, achieving an absolute average improvement of 2.04%. These results confirm that simple feature fusion enhances VAD robustness while maintaining computational efficiency.
Overlapping Speech Detection (OSD) aims to identify regions where multiple speakers overlap in a conversation, a critical challenge in multi-party speech processing. This work proposes a speaker-aware progressive OSD model that leverages a progressive training strategy to enhance the correlation between subtasks such as voice activity detection (VAD) and overlap detection. To improve acoustic representation, we explore the effectiveness of state-of-the-art self-supervised learning (SSL) models, including WavLM and wav2vec 2.0, while incorporating a speaker attention module to enrich features with frame-level speaker information. Experimental results show that the proposed method achieves state-of-the-art performance, with an F1 score of 82.76\% on the AMI test set, demonstrating its robustness and effectiveness in OSD.
ASR endpointing (EP) plays a major role in delivering a good user experience in products supporting human or artificial agents in human-human/machine conversations. Transducer-based ASR (T-ASR) is an end-to-end (E2E) ASR modelling technique preferred for streaming. A major limitation of T-ASR is delayed emission of ASR outputs, which could lead to errors or delays in EP. Inaccurate EP will cut the user off while speaking, returning incomplete transcript while delays in EP will increase the perceived latency, degrading the user experience. We propose methods to improve EP by addressing delayed emission along with EP mistakes. To address the delayed emission problem, we introduce an end-of-word token at the end of each word, along with a delay penalty. The EP delay is addressed by obtaining a reliable frame-level speech activity detection using an auxiliary network. We apply the proposed methods on Switchboard conversational speech corpus and evaluate it against a delay penalty method.
Massive machine-type communications (mMTC) are fundamental to the Internet of Things (IoT) framework in future wireless networks, involving the connection of a vast number of devices with sporadic transmission patterns. Traditional device activity detection (AD) methods are typically developed for Gaussian noise, but their performance may deteriorate when these conditions are not met, particularly in the presence of heavy-tailed impulsive noise. In this paper, we propose robust statistical techniques for AD that do not rely on the Gaussian assumption and replace the Gaussian loss function with robust loss functions that can effectively mitigate the impact of heavy-tailed noise and outliers. First, we prove that the coordinate-wise (conditional) objective function is geodesically convex and derive a fixed-point (FP) algorithm for minimizing it, along with convergence guarantees. Building on the FP algorithm, we propose two robust algorithms for solving the full (unconditional) objective function: a coordinate-wise optimization algorithm (RCWO) and a greedy covariance learning-based matching pursuit algorithm (RCL-MP). Numerical experiments demonstrate that the proposed methods significantly outperform existing algorithms in scenarios with non-Gaussian noise, achieving higher detection accuracy and robustness.
Speaker diarization, the process of identifying "who spoke when" in audio recordings, is essential for understanding classroom dynamics. However, classroom settings present distinct challenges, including poor recording quality, high levels of background noise, overlapping speech, and the difficulty of accurately capturing children's voices. This study investigates the effectiveness of multi-stage diarization models using Nvidia's NeMo diarization pipeline. We assess the impact of denoising on diarization accuracy and compare various voice activity detection (VAD) models, including self-supervised transformer-based frame-wise VAD models. We also explore a hybrid VAD approach that integrates Automatic Speech Recognition (ASR) word-level timestamps with frame-level VAD predictions. We conduct experiments using two datasets from English speaking classrooms to separate teacher vs. student speech and to separate all speakers. Our results show that denoising significantly improves the Diarization Error Rate (DER) by reducing the rate of missed speech. Additionally, training on both denoised and noisy datasets leads to substantial performance gains in noisy conditions. The hybrid VAD model leads to further improvements in speech detection, achieving a DER as low as 17% in teacher-student experiments and 45% in all-speaker experiments. However, we also identified trade-offs between voice activity detection and speaker confusion. Overall, our study highlights the effectiveness of multi-stage diarization models and integrating ASR-based information for enhancing speaker diarization in noisy classroom environments.
State-of-the-art spatio-temporal action detection (STAD) methods show promising results for extracting soccer events from broadcast videos. However, when operated in the high-recall, low-precision regime required for exhaustive event coverage in soccer analytics, their lack of contextual understanding becomes apparent: many false positives could be resolved by considering a broader sequence of actions and game-state information. In this work, we address this limitation by reasoning at the game level and improving STAD through the addition of a denoising sequence transduction task. Sequences of noisy, context-free player-centric predictions are processed alongside clean game state information using a Transformer-based encoder-decoder model. By modeling extended temporal context and reasoning jointly over team-level dynamics, our method leverages the "language of soccer" - its tactical regularities and inter-player dependencies - to generate "denoised" sequences of actions. This approach improves both precision and recall in low-confidence regimes, enabling more reliable event extraction from broadcast video and complementing existing pixel-based methods.
In this paper, we examine a key limitation in query-based detectors for temporal action detection (TAD), which arises from their direct adaptation of originally designed architectures for object detection. Despite the effectiveness of the existing models, they struggle to fully address the unique challenges of TAD, such as the redundancy in multi-scale features and the limited ability to capture sufficient temporal context. To address these issues, we propose a multi-dilated gated encoder and central-adjacent region integrated decoder for temporal action detection transformer (DiGIT). Our approach replaces the existing encoder that consists of multi-scale deformable attention and feedforward network with our multi-dilated gated encoder. Our proposed encoder reduces the redundant information caused by multi-level features while maintaining the ability to capture fine-grained and long-range temporal information. Furthermore, we introduce a central-adjacent region integrated decoder that leverages a more comprehensive sampling strategy for deformable cross-attention to capture the essential information. Extensive experiments demonstrate that DiGIT achieves state-of-the-art performance on THUMOS14, ActivityNet v1.3, and HACS-Segment. Code is available at: https://github.com/Dotori-HJ/DiGIT
Integrated sensing, communication, and computation (ISCC) has been regarded as a prospective technology for the next-generation wireless network, supporting humancentric intelligent applications. However, the delay sensitivity of these computation-intensive applications, especially in a multidevice ISCC system with limited resources, highlights the urgent need for efficient sensing task execution frameworks. To address this, we propose a resource-efficient sensing framework in this paper. Different from existing solutions, it features a novel action detection module deployed at each device to detect the onset of an action. Only time windows filled with signals of interest are offloaded to the edge server and processed by the edge recognition module, thus reducing overhead. Furthermore, we quantitatively analyze the sensing performance of the proposed sensing framework and formulate a sensing accuracy maximization problem under power, delay, and resource limitations for the multi-device ISCC system. By decomposing it into two subproblems, we develop an alternating direction method of multipliers (ADMM)-based distributed algorithm. It alternatively solves a sensing accuracy maximization subproblem at each device and employs a closed-form computation resource allocation strategy at the edge server till convergence. Finally, a real-world test is conducted using commodity wireless devices to validate the sensing performance analysis. Extensive test results demonstrate that our proposal achieves higher sensing accuracy under the limited resource compared to two baselines.
Temporal Action Detection and Moment Retrieval constitute two pivotal tasks in video understanding, focusing on precisely localizing temporal segments corresponding to specific actions or events. Recent advancements introduced Moment Detection to unify these two tasks, yet existing approaches remain confined to closed-set scenarios, limiting their applicability in open-world contexts. To bridge this gap, we present Grounding-MD, an innovative, grounded video-language pre-training framework tailored for open-world moment detection. Our framework incorporates an arbitrary number of open-ended natural language queries through a structured prompt mechanism, enabling flexible and scalable moment detection. Grounding-MD leverages a Cross-Modality Fusion Encoder and a Text-Guided Fusion Decoder to facilitate comprehensive video-text alignment and enable effective cross-task collaboration. Through large-scale pre-training on temporal action detection and moment retrieval datasets, Grounding-MD demonstrates exceptional semantic representation learning capabilities, effectively handling diverse and complex query conditions. Comprehensive evaluations across four benchmark datasets including ActivityNet, THUMOS14, ActivityNet-Captions, and Charades-STA demonstrate that Grounding-MD establishes new state-of-the-art performance in zero-shot and supervised settings in open-world moment detection scenarios. All source code and trained models will be released.
This work presents MicroNAS, an automated neural architecture search tool specifically designed to create models optimized for microcontrollers with small memory resources. The ESP32 microcontroller, with 320 KB of memory, is used as the target platform. The artificial intelligence contribution lies in a novel method for optimizing convolutional neural network and gated recurrent unit architectures by considering the memory size of the target microcontroller as a guide. A comparison is made between memory-driven model optimization and traditional two-stage methods, which use pruning, to show the effectiveness of the proposed framework. To demonstrate the engineering application of MicroNAS, a fall detection system (FDS) for lower-limb amputees is developed as a pilot study. A critical challenge in fall detection studies, class imbalance in the dataset, is addressed. The results show that MicroNAS models achieved higher F1-scores than alternative approaches, such as ensemble methods and H2O Automated Machine Learning, presenting a significant step forward in real-time FDS development. Biomechanists using body-worn sensors for activity detection can adopt the open-source code to design machine learning models tailored for microcontroller platforms with limited memory.
In this work, we focus on scaling open-vocabulary action detection. Existing approaches for action detection are predominantly limited to closed-set scenarios and rely on complex, parameter-heavy architectures. Extending these models to the open-vocabulary setting poses two key challenges: (1) the lack of large-scale datasets with many action classes for robust training, and (2) parameter-heavy adaptations to a pretrained vision-language contrastive model to convert it for detection, risking overfitting the additional non-pretrained parameters to base action classes. Firstly, we introduce an encoder-only multimodal model for video action detection, reducing the reliance on parameter-heavy additions for video action detection. Secondly, we introduce a simple weakly supervised training strategy to exploit an existing closed-set action detection dataset for pretraining. Finally, we depart from the ill-posed base-to-novel benchmark used by prior works in open-vocabulary action detection and devise a new benchmark to evaluate on existing closed-set action detection datasets without ever using them for training, showing novel results to serve as baselines for future work.
Temporal action detection aims to locate and classify actions in untrimmed videos. While recent works focus on designing powerful feature processors for pre-trained representations, they often overlook the inherent noise and redundancy within these features. Large-scale pre-trained video encoders tend to introduce background clutter and irrelevant semantics, leading to context confusion and imprecise boundaries. To address this, we propose a frequency-aware decoupling network that improves action discriminability by filtering out noisy semantics captured by pre-trained models. Specifically, we introduce an adaptive temporal decoupling scheme that suppresses irrelevant information while preserving fine-grained atomic action details, yielding more task-specific representations. In addition, we enhance inter-frame modeling by capturing temporal variations to better distinguish actions from background redundancy. Furthermore, we present a long-short-term category-aware relation network that jointly models local transitions and long-range dependencies, improving localization precision. The refined atomic features and frequency-guided dynamics are fed into a standard detection head to produce accurate action predictions. Extensive experiments on THUMOS14, HACS, and ActivityNet-1.3 show that our method, powered by InternVideo2-6B features, achieves state-of-the-art performance on temporal action detection benchmarks.
Temporal action detection (TAD) aims to identify and localize action instances in untrimmed videos, which is essential for various video understanding tasks. However, recent improvements in model performance, driven by larger feature extractors and datasets, have led to increased computational demands. This presents a challenge for applications like autonomous driving and robotics, which rely on limited computational resources. While existing channel pruning methods can compress these models, reducing the number of channels often hinders the parallelization efficiency of GPU, due to the inefficient multiplication between small matrices. Instead of pruning channels, we propose a Progressive Block Drop method that reduces model depth while retaining layer width. In this way, we still use large matrices for computation but reduce the number of multiplications. Our approach iteratively removes redundant blocks in two steps: first, we drop blocks with minimal impact on model performance; and second, we employ a parameter-efficient cross-depth alignment technique, fine-tuning the pruned model to restore model accuracy. Our method achieves a 25% reduction in computational overhead on two TAD benchmarks (THUMOS14 and ActivityNet-1.3) to achieve lossless compression. More critically, we empirically show that our method is orthogonal to channel pruning methods and can be combined with it to yield further efficiency gains.
Online Action Detection (OAD) detects actions in streaming videos using past observations. State-of-the-art OAD approaches model past observations and their interactions with an anticipated future. The past is encoded using short- and long-term memories to capture immediate and long-range dependencies, while anticipation compensates for missing future context. We identify a training-inference discrepancy in existing OAD methods that hinders learning effectiveness. The training uses varying lengths of short-term memory, while inference relies on a full-length short-term memory. As a remedy, we propose a Context-enhanced Memory-Refined Transformer (CMeRT). CMeRT introduces a context-enhanced encoder to improve frame representations using additional near-past context. It also features a memory-refined decoder to leverage near-future generation to enhance performance. CMeRT achieves state-of-the-art in online detection and anticipation on THUMOS'14, CrossTask, and EPIC-Kitchens-100.
Fast and accurate device activity detection is the critical challenge in grant-free access for supporting massive machine-type communications (mMTC) and ultra-reliable low-latency communications (URLLC) in 5G and beyond. The state-of-the-art methods have unsatisfactory error rates or computation times. To address these outstanding issues, we propose new maximum likelihood estimation (MLE) and maximum a posterior estimation (MAPE) based device activity detection methods for known and unknown pathloss that achieve superior error rate and computation time tradeoffs using optimization and deep learning techniques. Specifically, we investigate four non-convex optimization problems for MLE and MAPE in the two pathloss cases, with one MAPE problem being formulated for the first time. For each non-convex problem, we develop an innovative parallel iterative algorithm using the parallel successive convex approximation (PSCA) method. Each PSCA-based algorithm allows parallel computations, uses up to the objective function's second-order information, converges to the problem's stationary points, and has a low per-iteration computational complexity compared to the state-of-the-art algorithms. Then, for each PSCA-based iterative algorithm, we present a deep unrolling neural network implementation, called PSCA-Net, to further reduce the computation time. Each PSCA-Net elegantly marries the underlying PSCA-based algorithm's parallel computation mechanism with the parallelizable neural network architecture and effectively optimizes its step sizes based on vast data samples to speed up the convergence. Numerical results demonstrate that the proposed methods can significantly reduce the error rate and computation time compared to the state-of-the-art methods, revealing their significant values for grant-free access.