Search

Recently, end-to-end automatic speech recognition has become the mainstream approach in both industry and academia. To optimize system performance in specific scenarios, the Weighted Finite-State Transducer (WFST) is extensively used to integrate acoustic and language models, leveraging its capacity to implicitly fuse language models within static graphs, thereby ensuring robust recognition while also facilitating rapid error correction. However, WFST necessitates a frame-by-frame search of CTC posterior probabilities through autoregression, which significantly hampers inference speed. In this work, we thoroughly investigate the spike property of CTC outputs and further propose the conjecture that adjacent frames to non-blank spikes carry semantic information beneficial to the model. Building on this, we propose the Spike Window Decoding algorithm, which greatly improves the inference speed by making the number of frames decoded in WFST linearly related to the number of spiking frames in the CTC output, while guaranteeing the recognition performance. Our method achieves SOTA recognition accuracy with significantly accelerates decoding speed, proven across both AISHELL-1 and large-scale In-House datasets, establishing a pioneering approach for integrating CTC output with WFST., Comment: Accepted by ICASSP 2025

Humans can perceive speakers' characteristics (e.g., identity, gender, personality and emotion) by their appearance, which are generally aligned to their voice style. Recently, vision-driven Text-to-speech (TTS) scholars grounded their investigations on real-person faces, thereby restricting effective speech synthesis from applying to vast potential usage scenarios with diverse characters and image styles. To solve this issue, we introduce a novel FaceSpeak approach. It extracts salient identity characteristics and emotional representations from a wide variety of image styles. Meanwhile, it mitigates the extraneous information (e.g., background, clothing, and hair color, etc.), resulting in synthesized speech closely aligned with a character's persona. Furthermore, to overcome the scarcity of multi-modal TTS data, we have devised an innovative dataset, namely Expressive Multi-Modal TTS, which is diligently curated and annotated to facilitate research in this domain. The experimental results demonstrate our proposed FaceSpeak can generate portrait-aligned voice with satisfactory naturalness and quality.

Controlling the style and characteristics of speech synthesis is crucial for adapting the output to specific contexts and user requirements. Previous Text-to-speech (TTS) works have focused primarily on the technical aspects of producing natural-sounding speech, such as intonation, rhythm, and clarity. However, they overlook the fact that there is a growing emphasis on spatial perception of synthesized speech, which may provide immersive experience in gaming and virtual reality. To solve this issue, in this paper, we present a novel multi-modal TTS approach, namely Image-indicated Immersive Text-to-speech Synthesis (I2TTS). Specifically, we introduce a scene prompt encoder that integrates visual scene prompts directly into the synthesis pipeline to control the speech generation process. Additionally, we propose a reverberation classification and refinement technique that adjusts the synthesized mel-spectrogram to enhance the immersive experience, ensuring that the involved reverberation condition matches the scene accurately. Experimental results demonstrate that our model achieves high-quality scene and spatial matching without compromising speech naturalness, marking a significant advancement in the field of context-aware speech synthesis. Project demo page: https://spatialTTS.github.io/ Index Terms-Speech synthesis, scene prompt, spatial perception, Comment: The paper is missing some information

TSE(Target Speaker Extraction) aims to extract the clean speech of the target speaker in an audio mixture, thus eliminating irrelevant background noise and speech. While prior work has explored various auxiliary cues including pre-recorded speech, visual information (e.g., lip motions and gestures), and spatial information, the acquisition and selection of such strong cues are infeasible in many practical scenarios. Unlike all existing work, in this paper, we condition the TSE algorithm on semantic cues extracted from limited and unaligned text content, such as condensed points from a presentation slide. This method is particularly useful in scenarios like meetings, poster sessions, or lecture presentations, where acquiring other cues in real-time is challenging. To this end, we design two different networks. Specifically, our proposed TPE fuses audio features with content-based semantic cues to facilitate time-frequency mask generation to filter out extraneous noise, while another proposal, namely TSR, employs the contrastive learning technique to associate blindly separated speech signals with semantic cues. The experimental results show the efficacy in accurately identifying the target speaker by utilizing semantic cues derived from limited and unaligned text, resulting in SI-SDRi of 12.16 dB, SDRi of 12.66 dB, PESQi of 0.830 and STOIi of 0.150, respectively. Dataset and source code will be publicly available. Project demo page: https://slideTSE.github.io/.

Open-vocabulary object detection (OVD) models are considered to be Large Multi-modal Models (LMM), due to their extensive training data and a large number of parameters. Mainstream OVD models prioritize object coarse-grained category rather than focus on their fine-grained attributes, e.g., colors or materials, thus failed to identify objects specified with certain attributes. However, OVD models are pretrained on large-scale image-text pairs with rich attribute words, whose latent feature space can represent the global text feature as a linear composition of fine-grained attribute tokens without highlighting them. Therefore, we propose in this paper a universal and explicit approach for frozen mainstream OVD models that boosts their attribute-level detection capabilities by highlighting fine-grained attributes in explicit linear space. Firstly, a LLM is leveraged to highlight attribute words within the input text as a zero-shot prompted task. Secondly, by strategically adjusting the token masks, the text encoders of OVD models extract both global text and attribute-specific features, which are then explicitly composited as two vectors in linear space to form the new attribute-highlighted feature for detection tasks, where corresponding scalars are hand-crafted or learned to reweight both two vectors. Notably, these scalars can be seamlessly transferred among different OVD models, which proves that such an explicit linear composition is universal. Empirical evaluation on the FG-OVD dataset demonstrates that our proposed method uniformly improves fine-grained attribute-level OVD of various mainstream models and achieves new state-of-the-art performance., Comment: This work has been submitted to the IEEE for possible publication

Task inharmony problem commonly occurs in modern object detectors, leading to inconsistent qualities between classification and regression tasks. The predicted boxes with high classification scores but poor localization positions or low classification scores but accurate localization positions will worsen the performance of detectors after Non-Maximum Suppression. Furthermore, when object detectors collaborate with Quantization-Aware Training (QAT), we observe that the task inharmony problem will be further exacerbated, which is considered one of the main causes of the performance degradation of quantized detectors. To tackle this issue, we propose the Harmonious Quantization for Object Detection (HQOD) framework, which consists of two components. Firstly, we propose a task-correlated loss to encourage detectors to focus on improving samples with lower task harmony quality during QAT. Secondly, a harmonious Intersection over Union (IoU) loss is incorporated to balance the optimization of the regression branch across different IoU levels. The proposed HQOD can be easily integrated into different QAT algorithms and detectors. Remarkably, on the MS COCO dataset, our 4-bit ATSS with ResNet-50 backbone achieves a state-of-the-art mAP of 39.6%, even surpassing the full-precision one., Comment: 2024 IEEE International Conference on Multimedia and Expo (ICME), July 15 - July 19, 2024, Niagra Falls, Ontario, Canada

Video-language alignment is a crucial multi-modal task that benefits various downstream applications, e.g., video-text retrieval and video question answering. Existing methods either utilize multi-modal information in video-text pairs or apply global and local alignment techniques to promote alignment precision. However, these methods often fail to fully explore the spatio-temporal relationships among vision tokens within video and across different video-text pairs. In this paper, we propose a novel Spatio-Temporal Graph Transformer module to uniformly learn spatial and temporal contexts for video-language alignment pre-training (dubbed STGT). Specifically, our STGT combines spatio-temporal graph structure information with attention in transformer block, effectively utilizing the spatio-temporal contexts. In this way, we can model the relationships between vision tokens, promoting video-text alignment precision for benefiting downstream tasks. In addition, we propose a self-similarity alignment loss to explore the inherent self-similarity in the video and text. With the initial optimization achieved by contrastive learning, it can further promote the alignment accuracy between video and text. Experimental results on challenging downstream tasks, including video-text retrieval and video question answering, verify the superior performance of our method., Comment: under review

Existing open-set recognition (OSR) studies typically assume that each image contains only one class label, with the unknown test set (negative) having a disjoint label space from the known test set (positive), a scenario referred to as full-label shift. This paper introduces the mixed OSR problem, where test images contain multiple class semantics, with both known and unknown classes co-occurring in the negatives, leading to a more complex super-label shift that better reflects real-world scenarios. To tackle this challenge, we propose the OpenSlot framework, based on object-centric learning, which uses slot features to represent diverse class semantics and generate class predictions. The proposed anti-noise slot (ANS) technique helps mitigate the impact of noise (invalid or background) slots during classification training, addressing the semantic misalignment between class predictions and ground truth. We evaluate OpenSlot on both mixed and conventional OSR benchmarks. Without elaborate designs, our method not only excels existing approaches in detecting super-label shifts across OSR tasks, but also achieves state-of-the-art performance on conventional benchmarks. Meanwhile, OpenSlot can localize class objects without using bounding boxes during training, demonstrating competitive performance in open-set object detection and potential for generalization., Comment: This study is under IEEE TMM review

Generating reliable pseudo masks from image-level labels is challenging in the weakly supervised semantic segmentation (WSSS) task due to the lack of spatial information. Prevalent class activation map (CAM)-based solutions are challenged to discriminate the foreground (FG) objects from the suspicious background (BG) pixels (a.k.a. co-occurring) and learn the integral object regions. This paper proposes a simple fine-grained background representation (FBR) method to discover and represent diverse BG semantics and address the co-occurring problems. We abandon using the class prototype or pixel-level features for BG representation. Instead, we develop a novel primitive, negative region of interest (NROI), to capture the fine-grained BG semantic information and conduct the pixel-to-NROI contrast to distinguish the confusing BG pixels. We also present an active sampling strategy to mine the FG negatives on-the-fly, enabling efficient pixel-to-pixel intra-foreground contrastive learning to activate the entire object region. Thanks to the simplicity of design and convenience in use, our proposed method can be seamlessly plugged into various models, yielding new state-of-the-art results under various WSSS settings across benchmarks. Leveraging solely image-level (I) labels as supervision, our method achieves 73.2 mIoU and 45.6 mIoU segmentation results on Pascal Voc and MS COCO test sets, respectively. Furthermore, by incorporating saliency maps as an additional supervision signal (I+S), we attain 74.9 mIoU on Pascal Voc test set. Concurrently, our FBR approach demonstrates meaningful performance gains in weakly-supervised instance segmentation (WSIS) tasks, showcasing its robustness and strong generalization capabilities across diverse domains.

Distinguished from traditional knowledge graphs (KGs), temporal knowledge graphs (TKGs) must explore and reason over temporally evolving facts adequately. However, existing TKG approaches still face two main challenges, i.e., the limited capability to model arbitrary timestamps continuously and the lack of rich inference patterns under temporal constraints. In this paper, we propose an innovative TKGE method (PTBox) via polynomial decomposition-based temporal representation and box embedding-based entity representation to tackle the above-mentioned problems. Specifically, we decompose time information by polynomials and then enhance the model's capability to represent arbitrary timestamps flexibly by incorporating the learnable temporal basis tensor. In addition, we model every entity as a hyperrectangle box and define each relation as a transformation on the head and tail entity boxes. The entity boxes can capture complex geometric structures and learn robust representations, improving the model's inductive capability for rich inference patterns. Theoretically, our PTBox can encode arbitrary time information or even unseen timestamps while capturing rich inference patterns and higher-arity relations of the knowledge base. Extensive experiments on real-world datasets demonstrate the effectiveness of our method., Comment: Accepted by LREC-COLING 2024 (long paper, camera-ready version)

Temporal Knowledge Graph (TKG) reasoning often involves completing missing factual elements along the timeline. Although existing methods can learn good embeddings for each factual element in quadruples by integrating temporal information, they often fail to infer the evolution of temporal facts. This is mainly because of (1) insufficiently exploring the internal structure and semantic relationships within individual quadruples and (2) inadequately learning a unified representation of the contextual and temporal correlations among different quadruples. To overcome these limitations, we propose a novel Transformer-based reasoning model (dubbed ECEformer) for TKG to learn the Evolutionary Chain of Events (ECE). Specifically, we unfold the neighborhood subgraph of an entity node in chronological order, forming an evolutionary chain of events as the input for our model. Subsequently, we utilize a Transformer encoder to learn the embeddings of intra-quadruples for ECE. We then craft a mixed-context reasoning module based on the multi-layer perceptron (MLP) to learn the unified representations of inter-quadruples for ECE while accomplishing temporal knowledge reasoning. In addition, to enhance the timeliness of the events, we devise an additional time prediction task to complete effective temporal information within the learned unified representation. Extensive experiments on six benchmark datasets verify the state-of-the-art performance and the effectiveness of our method., Comment: Accepted by SIGIR 2024 (the Full paper track, camera ready version)

Scene text spotting is a challenging task, especially for inverse-like scene text, which has complex layouts, e.g., mirrored, symmetrical, or retro-flexed. In this paper, we propose a unified end-to-end trainable inverse-like antagonistic text spotting framework dubbed IATS, which can effectively spot inverse-like scene texts without sacrificing general ones. Specifically, we propose an innovative reading-order estimation module (REM) that extracts reading-order information from the initial text boundary generated by an initial boundary module (IBM). To optimize and train REM, we propose a joint reading-order estimation loss consisting of a classification loss, an orthogonality loss, and a distribution loss. With the help of IBM, we can divide the initial text boundary into two symmetric control points and iteratively refine the new text boundary using a lightweight boundary refinement module (BRM) for adapting to various shapes and scales. To alleviate the incompatibility between text detection and recognition, we propose a dynamic sampling module (DSM) with a thin-plate spline that can dynamically sample appropriate features for recognition in the detected text region. Without extra supervision, the DSM can proactively learn to sample appropriate features for text recognition through the gradient returned by the recognition module. Extensive experiments on both challenging scene text and inverse-like scene text datasets demonstrate that our method achieves superior performance both on irregular and inverse-like text spotting., Comment: 14 pages, 16 figures, Accepted by TIP-2024

In this work, we use multi-view aerial images to reconstruct the geometry, lighting, and material of facades using neural signed distance fields (SDFs). Without the requirement of complex equipment, our method only takes simple RGB images captured by a drone as inputs to enable physically based and photorealistic novel-view rendering, relighting, and editing. However, a real-world facade usually has complex appearances ranging from diffuse rocks with subtle details to large-area glass windows with specular reflections, making it hard to attend to everything. As a result, previous methods can preserve the geometry details but fail to reconstruct smooth glass windows or verse vise. In order to address this challenge, we introduce three spatial- and semantic-adaptive optimization strategies, including a semantic regularization approach based on zero-shot segmentation techniques to improve material consistency, a frequency-aware geometry regularization to balance surface smoothness and details in different surfaces, and a visibility probe-based scheme to enable efficient modeling of the local lighting in large-scale outdoor environments. In addition, we capture a real-world facade aerial 3D scanning image set and corresponding point clouds for training and benchmarking. The experiment demonstrates the superior quality of our method on facade holistic inverse rendering, novel view synthesis, and scene editing compared to state-of-the-art baselines.

Optical quantum sensing promises measurement precision beyond classical sensors termed the Heisenberg limit (HL). However, conventional methodologies often rely on prior knowledge of the target system to achieve HL, presenting challenges in practical applications. Addressing this limitation, we introduce an innovative Deep Learning-based Quantum Sensing scheme (DQS), enabling optical quantum sensors to attain HL in agnostic environments. DQS incorporates two essential components: a Graph Neural Network (GNN) predictor and a trigonometric interpolation algorithm. Operating within a data-driven paradigm, DQS utilizes the GNN predictor, trained on offline data, to unveil the intrinsic relationships between the optical setups employed in preparing the probe state and the resulting quantum Fisher information (QFI) after interaction with the agnostic environment. This distilled knowledge facilitates the identification of optimal optical setups associated with maximal QFI. Subsequently, DQS employs a trigonometric interpolation algorithm to recover the unknown parameter estimates for the identified optical setups. Extensive experiments are conducted to investigate the performance of DQS under different settings up to eight photons. Our findings not only offer a new lens through which to accelerate optical quantum sensing tasks but also catalyze future research integrating deep learning and quantum mechanics.

Unsupervised domain adaptation (UDA) has been a potent technique to handle the lack of annotations in the target domain, particularly in semantic segmentation task. This study introduces a different UDA scenarios where the target domain contains unlabeled video frames. Drawing upon recent advancements of self-supervised learning of the object motion from unlabeled videos with geometric constraint, we design a \textbf{Mo}tion-guided \textbf{D}omain \textbf{A}daptive semantic segmentation framework (MoDA). MoDA harnesses the self-supervised object motion cues to facilitate cross-domain alignment for segmentation task. First, we present an object discovery module to localize and segment target moving objects using object motion information. Then, we propose a semantic mining module that takes the object masks to refine the pseudo labels in the target domain. Subsequently, these high-quality pseudo labels are used in the self-training loop to bridge the cross-domain gap. On domain adaptive video and image segmentation experiments, MoDA shows the effectiveness utilizing object motion as guidance for domain alignment compared with optical flow information. Moreover, MoDA exhibits versatility as it can complement existing state-of-the-art UDA approaches. Code at https://github.com/feipanir/MoDA., Comment: CVPR 2024 Workshop on Learning with Limited Labelled Data for Image and Video Understanding. Best Paper Award

The local and global features are both essential for automatic speech recognition (ASR). Many recent methods have verified that simply combining local and global features can further promote ASR performance. However, these methods pay less attention to the interaction of local and global features, and their series architectures are rigid to reflect local and global relationships. To address these issues, this paper proposes InterFormer for interactive local and global features fusion to learn a better representation for ASR. Specifically, we combine the convolution block with the transformer block in a parallel design. Besides, we propose a bidirectional feature interaction module (BFIM) and a selective fusion module (SFM) to implement the interaction and fusion of local and global features, respectively. Extensive experiments on public ASR datasets demonstrate the effectiveness of our proposed InterFormer and its superior performance over the other Transformer and Conformer models., Comment: Accepted by Interspeech 2023

Attention-based encoder-decoder (AED) models have shown impressive performance in ASR. However, most existing AED methods neglect to simultaneously leverage both acoustic and semantic features in decoder, which is crucial for generating more accurate and informative semantic states. In this paper, we propose an Acoustic and Semantic Cooperative Decoder (ASCD) for ASR. In particular, unlike vanilla decoders that process acoustic and semantic features in two separate stages, ASCD integrates them cooperatively. To prevent information leakage during training, we design a Causal Multimodal Mask. Moreover, a variant Semi-ASCD is proposed to balance accuracy and computational cost. Our proposal is evaluated on the publicly available AISHELL-1 and aidatatang_200zh datasets using Transformer, Conformer, and Branchformer as encoders, respectively. The experimental results show that ASCD significantly improves the performance by leveraging both the acoustic and semantic information cooperatively., Comment: Accepted by Interspeech 2023

With the prosperity of the video surveillance, multiple cameras have been applied to accurately locate pedestrians in a specific area. However, previous methods rely on the human-labeled annotations in every video frame and camera view, leading to heavier burden than necessary camera calibration and synchronization. Therefore, we propose in this paper an Unsupervised Multi-view Pedestrian Detection approach (UMPD) to eliminate the need of annotations to learn a multi-view pedestrian detector via 2D-3D mapping. 1) Firstly, Semantic-aware Iterative Segmentation (SIS) is proposed to extract unsupervised representations of multi-view images, which are converted into 2D pedestrian masks as pseudo labels, via our proposed iterative PCA and zero-shot semantic classes from vision-language models. 2) Secondly, we propose Geometry-aware Volume-based Detector (GVD) to end-to-end encode multi-view 2D images into a 3D volume to predict voxel-wise density and color via 2D-to-3D geometric projection, trained by 3D-to-2D rendering losses with SIS pseudo labels. 3) Thirdly, for better detection results, i.e., the 3D density projected on Birds-Eye-View from GVD, we propose Vertical-aware BEV Regularization (VBR) to constraint them to be vertical like the natural pedestrian poses. Extensive experiments on popular multi-view pedestrian detection benchmarks Wildtrack, Terrace, and MultiviewX, show that our proposed UMPD approach, as the first fully-unsupervised method to our best knowledge, performs competitively to the previous state-of-the-art supervised techniques. Code will be available.

Detecting pedestrians accurately in urban scenes is significant for realistic applications like autonomous driving or video surveillance. However, confusing human-like objects often lead to wrong detections, and small scale or heavily occluded pedestrians are easily missed due to their unusual appearances. To address these challenges, only object regions are inadequate, thus how to fully utilize more explicit and semantic contexts becomes a key problem. Meanwhile, previous context-aware pedestrian detectors either only learn latent contexts with visual clues, or need laborious annotations to obtain explicit and semantic contexts. Therefore, we propose in this paper a novel approach via Vision-Language semantic self-supervision for context-aware Pedestrian Detection (VLPD) to model explicitly semantic contexts without any extra annotations. Firstly, we propose a self-supervised Vision-Language Semantic (VLS) segmentation method, which learns both fully-supervised pedestrian detection and contextual segmentation via self-generated explicit labels of semantic classes by vision-language models. Furthermore, a self-supervised Prototypical Semantic Contrastive (PSC) learning method is proposed to better discriminate pedestrians and other classes, based on more explicit and semantic contexts obtained from VLS. Extensive experiments on popular benchmarks show that our proposed VLPD achieves superior performances over the previous state-of-the-arts, particularly under challenging circumstances like small scale and heavy occlusion. Code is available at https://github.com/lmy98129/VLPD., Comment: Accepted by CVPR 2023