Text-based generation and editing of 3D scenes hold significant potential for streamlining content creation through intuitive user interactions. While recent advances leverage 3D Gaussian Splatting (3DGS) for high-fidelity and real-time rendering, existing methods are often specialized and task-focused, lacking a unified framework for both generation and editing. In this paper, we introduce SplatFlow, a comprehensive framework that addresses this gap by enabling direct 3DGS generation and editing. SplatFlow comprises two main components: a multi-view rectified flow (RF) model and a Gaussian Splatting Decoder (GSDecoder). The multi-view RF model operates in latent space, generating multi-view images, depths, and camera poses simultaneously, conditioned on text prompts, thus addressing challenges like diverse scene scales and complex camera trajectories in real-world settings. Then, the GSDecoder efficiently translates these latent outputs into 3DGS representations through a feed-forward 3DGS method. Leveraging training-free inversion and inpainting techniques, SplatFlow enables seamless 3DGS editing and supports a broad range of 3D tasks-including object editing, novel view synthesis, and camera pose estimation-within a unified framework without requiring additional complex pipelines. We validate SplatFlow's capabilities on the MVImgNet and DL3DV-7K datasets, demonstrating its versatility and effectiveness in various 3D generation, editing, and inpainting-based tasks.
Transformer-based pre-trained language models have dominated the field of Natural Language Processing (NLP) for quite some time now. However, the Nepali language, spoken by approximately 32 million people worldwide, remains significantly underrepresented in this domain. This underrepresentation is primarily attributed to the scarcity of monolingual data corpora and limited available resources for the Nepali language. While existing efforts have predominantly concentrated on basic encoder-based models, there is a notable gap in the exploration of decoder-based architectures. To address this gap, we have collected 27.5 GB of Nepali text data, approximately 2.4x larger than any previously available Nepali language corpus. Leveraging this data, we pre-trained three different models i.e., BERT, RoBERTa, and GPT-2, exclusively for the Nepali Language. Furthermore, we performed instruction tuning and explored its potential for monolingual Nepali data, providing a foundation for future research. Our models outperformed the existing best model by 2 points on Nep-gLUE benchmark, scoring 95.60 and also outperformed existing models on text generation tasks, demonstrating improvements in both understanding and generating Nepali text.
Autoregressive language models like GPT aim at predicting next tokens, while autoencoding models such as BERT are trained on tasks such as predicting masked tokens. We train a decoder only architecture for predicting the second last token for a sequence of tokens. Our approach yields higher computational training efficiency than BERT-style models by employing a structured deterministic approach towards masking tokens. We use our model to improve the next token predictions of a standard GPT by combining both predictions in a ``generate-then-refine'' approach. We show on different variants of GPT-2 and different datasets that (not unexpectedly) second last token predictions are much more accurate, i.e., more than 15\% higher accuracy than ordinary next token predictors. The ``generate-then-refine'' approach also demonstrates notable improvements in next-token predictions, yielding smaller yet consistent and significant gains.
Self-supervised learning (SSL) is pervasively exploited in training high-quality upstream encoders with a large amount of unlabeled data. However, it is found to be susceptible to backdoor attacks merely via polluting a small portion of training data. The victim encoders associate triggered inputs with target embeddings, e.g., mapping a triggered cat image to an airplane embedding, such that the downstream tasks inherit unintended behaviors when the trigger is activated. Emerging backdoor attacks have shown great threats across different SSL paradigms such as contrastive learning and CLIP, yet limited research is devoted to defending against such attacks, and existing defenses fall short in detecting advanced stealthy backdoors. To address the limitations, we propose a novel detection mechanism, DeDe, which detects the activation of backdoor mappings caused by triggered inputs on victim encoders. Specifically, DeDe trains a decoder for any given SSL encoder using an auxiliary dataset (which can be out-of-distribution or even slightly poisoned), so that for any triggered input that misleads the encoder into the target embedding, the decoder generates an output image significantly different from the input. DeDe leverages the discrepancy between the input and the decoded output to identify potential backdoor misbehavior during inference. We empirically evaluate DeDe on both contrastive learning and CLIP models against various types of backdoor attacks. Our results demonstrate promising detection effectiveness over various advanced attacks and superior performance compared over state-of-the-art detection methods.
In the domain of single-view 3D reconstruction, traditional techniques have frequently relied on expensive and time-intensive 3D annotation data. Facing the challenge of annotation acquisition, semi-supervised learning strategies offer an innovative approach to reduce the dependence on labeled data. Despite these developments, the utilization of this learning paradigm in 3D reconstruction tasks remains relatively constrained. In this research, we created an innovative semi-supervised framework for 3D reconstruction that distinctively uniquely introduces a multi shape prior fusion strategy, intending to guide the creation of more realistic object structures. Additionally, to improve the quality of shape generation, we integrated a self-attention module into the traditional decoder. In benchmark tests on the ShapeNet dataset, our method substantially outperformed existing supervised learning methods at diverse labeled ratios of 1\%, 10\%, and 20\%. Moreover, it showcased excellent performance on the real-world Pix3D dataset. Through comprehensive experiments on ShapeNet, our framework demonstrated a 3.3\% performance improvement over the baseline. Moreover, stringent ablation studies further confirmed the notable effectiveness of our approach. Our code has been released on https://github.com/NWUzhouwei/SSMP
Connectionist temporal classification (CTC)-based scene text recognition (STR) methods, e.g., SVTR, are widely employed in OCR applications, mainly due to their simple architecture, which only contains a visual model and a CTC-aligned linear classifier, and therefore fast inference. However, they generally have worse accuracy than encoder-decoder-based methods (EDTRs), particularly in challenging scenarios. In this paper, we propose SVTRv2, a CTC model that beats leading EDTRs in both accuracy and inference speed. SVTRv2 introduces novel upgrades to handle text irregularity and utilize linguistic context, which endows it with the capability to deal with challenging and diverse text instances. First, a multi-size resizing (MSR) strategy is proposed to adaptively resize the text and maintain its readability. Meanwhile, we introduce a feature rearrangement module (FRM) to ensure that visual features accommodate the alignment requirement of CTC well, thus alleviating the alignment puzzle. Second, we propose a semantic guidance module (SGM). It integrates linguistic context into the visual model, allowing it to leverage language information for improved accuracy. Moreover, SGM can be omitted at the inference stage and would not increase the inference cost. We evaluate SVTRv2 in both standard and recent challenging benchmarks, where SVTRv2 is fairly compared with 24 mainstream STR models across multiple scenarios, including different types of text irregularity, languages, and long text. The results indicate that SVTRv2 surpasses all the EDTRs across the scenarios in terms of accuracy and speed. Code is available at https://github.com/Topdu/OpenOCR.
Masked image modeling has achieved great success in learning representations but is limited by the huge computational costs. One cost-saving strategy makes the decoder reconstruct only a subset of masked tokens and throw the others, and we refer to this method as partial reconstruction. However, it also degrades the representation quality. Previous methods mitigate this issue by throwing tokens with minimal information using temporal redundancy inaccessible for static images or attention maps that incur extra costs and complexity. To address these limitations, we propose a progressive reconstruction strategy and a furthest sampling strategy to reconstruct those thrown tokens in an extremely lightweight way instead of completely abandoning them. This approach involves all masked tokens in supervision to ensure adequate pre-training, while maintaining the cost-reduction benefits of partial reconstruction. We validate the effectiveness of the proposed method across various existing frameworks. For example, when throwing 50% patches, we can achieve lossless performance of the ViT-B/16 while saving 28% FLOPs and 36% memory usage compared to standard MAE. Our source code will be made publicly available
Automated lumbar spine segmentation is very crucial for modern diagnosis systems. In this study, we introduce a novel machine-agnostic approach for segmenting lumbar vertebrae and intervertebral discs from MRI images, employing a cascaded model that synergizes an ROI detection and a Self-organized Operational Neural Network (Self-ONN)-based encoder-decoder network for segmentation. Addressing the challenge of diverse MRI modalities, our methodology capitalizes on a unique dataset comprising images from 12 scanners and 34 subjects, enhanced through strategic preprocessing and data augmentation techniques. The YOLOv8 medium model excels in ROI extraction, achieving an excellent performance of 0.916 mAP score. Significantly, our Self-ONN-based model, combined with a DenseNet121 encoder, demonstrates excellent performance in lumbar vertebrae and IVD segmentation with a mean Intersection over Union (IoU) of 83.66%, a sensitivity of 91.44%, and Dice Similarity Coefficient (DSC) of 91.03%, as validated through rigorous 10-fold cross-validation. This study not only showcases an effective approach to MRI segmentation in spine-related disorders but also sets the stage for future advancements in automated diagnostic tools, emphasizing the need for further dataset expansion and model refinement for broader clinical applicability.
Generating high-quality speech efficiently remains a key challenge for generative models in speech synthesis. This paper introduces VQalAttent, a lightweight model designed to generate fake speech with tunable performance and interpretability. Leveraging the AudioMNIST dataset, consisting of human utterances of decimal digits (0-9), our method employs a two-step architecture: first, a scalable vector quantized autoencoder (VQ-VAE) that compresses audio spectrograms into discrete latent representations, and second, a decoder-only transformer that learns the probability model of these latents. Trained transformer generates similar latent sequences, convertible to audio spectrograms by the VQ-VAE decoder, from which we generate fake utterances. Interpreting statistical and perceptual quality of the fakes, depending on the dimension and the extrinsic information of the latent space, enables guided improvements in larger, commercial generative models. As a valuable tool for understanding and refining audio synthesis, our results demonstrate VQalAttent's capacity to generate intelligible speech samples with limited computational resources, while the modularity and transparency of the training pipeline helps easily correlate the analytics with modular modifications, hence providing insights for the more complex models.
Existing Image Manipulation Localization (IML) methods mostly rely heavily on task-specific designs, making them perform well only on the target IML task, while joint training on multiple IML tasks causes significant performance degradation, hindering real applications. To this end, we propose Omni-IML, the first generalist model designed to unify IML across diverse tasks. Specifically, Omni-IML achieves generalization through three key components: (1) a Modal Gate Encoder, which adaptively selects the optimal encoding modality per sample, (2) a Dynamic Weight Decoder, which dynamically adjusts decoder filters to the task at hand, and (3) an Anomaly Enhancement module that leverages box supervision to highlight the tampered regions and facilitate the learning of task-agnostic features. Beyond localization, to support interpretation of the tampered images, we construct Omni-273k, a large high-quality dataset that includes natural language descriptions of tampered artifact. It is annotated through our automatic, chain-of-thoughts annotation technique. We also design a simple-yet-effective interpretation module to better utilize these descriptive annotations. Our extensive experiments show that our single Omni-IML model achieves state-of-the-art performance across all four major IML tasks, providing a valuable solution for practical deployment and a promising direction of generalist models in image forensics. Our code and dataset will be publicly available.
Open-vocabulary semantic segmentation aims to assign pixel-level labels to images across an unlimited range of classes. Traditional methods address this by sequentially connecting a powerful mask proposal generator, such as the Segment Anything Model (SAM), with a pre-trained vision-language model like CLIP. But these two-stage approaches often suffer from high computational costs, memory inefficiencies. In this paper, we propose ESC-Net, a novel one-stage open-vocabulary segmentation model that leverages the SAM decoder blocks for class-agnostic segmentation within an efficient inference framework. By embedding pseudo prompts generated from image-text correlations into SAM's promptable segmentation framework, ESC-Net achieves refined spatial aggregation for accurate mask predictions. ESC-Net achieves superior performance on standard benchmarks, including ADE20K, PASCAL-VOC, and PASCAL-Context, outperforming prior methods in both efficiency and accuracy. Comprehensive ablation studies further demonstrate its robustness across challenging conditions.
Video understanding has witnessed significant progress with recent video foundation models demonstrating strong performance owing to self-supervised pre-training objectives; Masked Autoencoders (MAE) being the design of choice. Nevertheless, the majority of prior works that leverage MAE pre-training have focused on relatively short video representations (16 / 32 frames in length) largely due to hardware memory and compute limitations that scale poorly with video length due to the dense memory-intensive self-attention decoding. One natural strategy to address these challenges is to subsample tokens to reconstruct during decoding (or decoder masking). In this work, we propose an effective strategy for prioritizing tokens which allows training on longer video sequences (128 frames) and gets better performance than, more typical, random and uniform masking strategies. The core of our approach is an adaptive decoder masking strategy that prioritizes the most important tokens and uses quantized tokens as reconstruction objectives. Our adaptive strategy leverages a powerful MAGVIT-based tokenizer that jointly learns the tokens and their priority. We validate our design choices through exhaustive ablations and observe improved performance of the resulting long-video (128 frames) encoders over short-video (32 frames) counterparts. With our long-video masked autoencoder (LVMAE) strategy, we surpass state-of-the-art on Diving48 by 3.9 points and EPIC-Kitchens-100 verb classification by 2.5 points while relying on a simple core architecture and video-only pre-training (unlike some of the prior works that require millions of labeled video-text pairs or specialized encoders).
We introduce a novel method for pre-training of large-scale vision encoders. Building on recent advancements in autoregressive pre-training of vision models, we extend this framework to a multimodal setting, i.e., images and text. In this paper, we present AIMV2, a family of generalist vision encoders characterized by a straightforward pre-training process, scalability, and remarkable performance across a range of downstream tasks. This is achieved by pairing the vision encoder with a multimodal decoder that autoregressively generates raw image patches and text tokens. Our encoders excel not only in multimodal evaluations but also in vision benchmarks such as localization, grounding, and classification. Notably, our AIMV2-3B encoder achieves 89.5% accuracy on ImageNet-1k with a frozen trunk. Furthermore, AIMV2 consistently outperforms state-of-the-art contrastive models (e.g., CLIP, SigLIP) in multimodal image understanding across diverse settings.
Multimodal large language models (MLLMs) are closing the gap to human visual perception capability rapidly, while, still lag behind on attending to subtle images details or locating small objects precisely, etc. Common schemes to tackle these issues include deploying multiple vision encoders or operating on original high-resolution images. Few studies have concentrated on taking the textual instruction into improving visual representation, resulting in losing focus in some vision-centric tasks, a phenomenon we herein termed as Amblyopia. In this work, we introduce Panther, a MLLM that closely adheres to user instruction and locates targets of interests precisely, with the finesse of a black panther. Specifically, Panther comprises three integral components: Panther-VE, Panther-Bridge, and Panther-Decoder. Panther-VE integrates user instruction information at the early stages of the vision encoder, thereby extracting the most relevant and useful visual representations. The Panther-Bridge module, equipped with powerful filtering capabilities, significantly reduces redundant visual information, leading to a substantial savings in training costs. The Panther-Decoder is versatile and can be employed with any decoder-only architecture of LLMs without discrimination. Experimental results, particularly on vision-centric benchmarks, have demonstrated the effectiveness of Panther.
Lossy compression methods rely on an autoencoder to transform a point cloud into latent points for storage, leaving the inherent redundancy of latent representations unexplored. To reduce redundancy in latent points, we propose a sparse priors guided method that achieves high reconstruction quality, especially at high compression ratios. This is accomplished by a dual-density scheme separately processing the latent points (intended for reconstruction) and the decoupled sparse priors (intended for storage). Our approach features an efficient dual-density data flow that relaxes size constraints on latent points, and hybridizes a progressive conditional diffusion model to encapsulate essential details for reconstruction within the conditions, which are decoupled hierarchically to intra-point and inter-point priors. Specifically, our method encodes the original point cloud into latent points and decoupled sparse priors through separate encoders. Latent points serve as intermediates, while sparse priors act as adaptive conditions. We then employ a progressive attention-based conditional denoiser to generate latent points conditioned on the decoupled priors, allowing the denoiser to dynamically attend to geometric and semantic cues from the priors at each encoding and decoding layer. Additionally, we integrate the local distribution into the arithmetic encoder and decoder to enhance local context modeling of the sparse points. The original point cloud is reconstructed through a point decoder. Compared to state-of-the-art, our method obtains superior rate-distortion trade-off, evidenced by extensive evaluations on the ShapeNet dataset and standard test datasets from MPEG group including 8iVFB, and Owlii.
This work is concerned with robust distributed multi-view image transmission over a severe fading channel with imperfect channel state information (CSI), wherein the sources are slightly correlated. Since the signals are further distorted at the decoder, traditional distributed deep joint source-channel coding (DJSCC) suffers considerable performance degradation. To tackle this problem, we leverage the complementarity and consistency characteristics among the distributed, yet correlated sources, and propose an enhanced robust DJSCC, namely RDJSCC. In RDJSCC, we design a novel cross-view information extraction (CVIE) mechanism to capture more nuanced cross-view patterns and dependencies. In addition, a complementarity-consistency fusion (CCF) mechanism is utilized to fuse the complementarity and consistency from multi-view information in a symmetric and compact manner. Theoretical analysis and simulation results show that our proposed RDJSCC can effectively leverage the advantages of correlated sources even under severe fading conditions, leading to an improved reconstruction performance. The open source code of this work is available at:https://dongbiao26.github.io/rdjscc/.
Medical image translation is crucial for reducing the need for redundant and expensive multi-modal imaging in clinical field. However, current approaches based on Convolutional Neural Networks (CNNs) and Transformers often fail to capture fine-grain semantic features, resulting in suboptimal image quality. To address this challenge, we propose SAM-I2I, a novel image-to-image translation framework based on the Segment Anything Model 2 (SAM2). SAM-I2I utilizes a pre-trained image encoder to extract multiscale semantic features from the source image and a decoder, based on the mask unit attention module, to synthesize target modality images. Our experiments on multi-contrast MRI datasets demonstrate that SAM-I2I outperforms state-of-the-art methods, offering more efficient and accurate medical image translation.
Connected and autonomous vehicles (CAVs) offload computationally intensive tasks to multi-access edge computing (MEC) servers via vehicle-to-infrastructure (V2I) communication, enabling applications within the vehicular metaverse, which transforms physical environment into the digital space enabling advanced analysis or predictive modeling. A core challenge is physical-to-virtual (P2V) synchronization through digital twins (DTs), reliant on MEC networks and ultra-reliable low-latency communication (URLLC). To address this, we introduce radiance field (RF) delta video compression (RFDVC), which uses RF-encoder and RF-decoder architecture using distributed RFs as DTs storing photorealistic 3D urban scenes in compressed form. This method extracts differences between CAV-frame capturing actual traffic and RF-frame capturing empty scene from the same camera pose in batches encoded and transmitted over the MEC network. Experiments show data savings up to 71% against H.264 codec and 44% against H.265 codec under different conditions as lighting changes, and rain. RFDVC also demonstrates resilience to transmission errors, achieving up to +0.29 structural similarity index measure (SSIM) improvement at block error rate (BLER) = 0.35 in non-rainy and +0.25 at BLER = 0.2 in rainy conditions, ensuring superior visual quality compared to standard video coding (VC) methods across various conditions.
The objective of Radiology Report Generation (RRG) is to automatically generate coherent textual analyses of diseases based on radiological images, thereby alleviating the workload of radiologists. Current AI-based methods for RRG primarily focus on modifications to the encoder-decoder model architecture. To advance these approaches, this paper introduces an Organ-Regional Information Driven (ORID) framework which can effectively integrate multi-modal information and reduce the influence of noise from unrelated organs. Specifically, based on the LLaVA-Med, we first construct an RRG-related instruction dataset to improve organ-regional diagnosis description ability and get the LLaVA-Med-RRG. After that, we propose an organ-based cross-modal fusion module to effectively combine the information from the organ-regional diagnosis description and radiology image. To further reduce the influence of noise from unrelated organs on the radiology report generation, we introduce an organ importance coefficient analysis module, which leverages Graph Neural Network (GNN) to examine the interconnections of the cross-modal information of each organ region. Extensive experiments an1d comparisons with state-of-the-art methods across various evaluation metrics demonstrate the superior performance of our proposed method.
We present MERLOT, a scalable mixture-of-expert (MoE) based refinement of distilled large language model optimized for encrypted traffic classification. By applying model distillation techniques in a teacher-student paradigm, compact models derived from GPT-2-base retain high classification accuracy while minimizing computational costs. These models function as specialized experts in an MoE architecture, dynamically assigned via a gating network. Unlike generation-based methods, our approach directly classifies encrypted traffic using the final decoder token with contextual feature embedding as input. Experiments on 10 datasets show superior or competitive performance over the state-of-the-art models while significantly reducing resource demands, underscoring its effectiveness and robustness.