Deepu Rajan

Heqing Zou, Yuke Si, Chen Chen et al.

Speech Emotion Recognition (SER) aims to help the machine to understand human's subjective emotion from only audio information. However, extracting and utilizing comprehensive in-depth audio information is still a challenging task. In this paper, we propose an end-to-end speech emotion recognition system using multi-level acoustic information with a newly designed co-attention module. We firstly extract multi-level acoustic information, including MFCC, spectrogram, and the embedded high-level acoustic information with CNN, BiLSTM and wav2vec2, respectively. Then these extracted features are treated as multimodal inputs and fused by the proposed co-attention mechanism. Experiments are carried on the IEMOCAP dataset, and our model achieves competitive performance with two different speaker-independent cross-validation strategies. Our code is available on GitHub.

Meng Shen, Yizheng Huang, Jianxiong Yin et al.

Training multimodal networks requires a vast amount of data due to their larger parameter space compared to unimodal networks. Active learning is a widely used technique for reducing data annotation costs by selecting only those samples that could contribute to improving model performance. However, current active learning strategies are mostly designed for unimodal tasks, and when applied to multimodal data, they often result in biased sample selection from the dominant modality. This unfairness hinders balanced multimodal learning, which is crucial for achieving optimal performance. To address this issue, we propose three guidelines for designing a more balanced multimodal active learning strategy. Following these guidelines, a novel approach is proposed to achieve more fair data selection by modulating the gradient embedding with the dominance degree among modalities. Our studies demonstrate that the proposed method achieves more balanced multimodal learning by avoiding greedy sample selection from the dominant modality. Our approach outperforms existing active learning strategies on a variety of multimodal classification tasks. Overall, our work highlights the importance of balancing sample selection in multimodal active learning and provides a practical solution for achieving more balanced active learning for multimodal classification.

Hui Lu, Yi Yu, Song Xia et al.

Large-scale Video Foundation Models (VFMs) has significantly advanced various video-related tasks, either through task-specific models or Multi-modal Large Language Models (MLLMs). However, the open accessibility of VFMs also introduces critical security risks, as adversaries can exploit full knowledge of the VFMs to launch potent attacks. This paper investigates a novel and practical adversarial threat scenario: attacking downstream models or MLLMs fine-tuned from open-source VFMs, without requiring access to the victim task, training data, model query, and architecture. In contrast to conventional transfer-based attacks that rely on task-aligned surrogate models, we demonstrate that adversarial vulnerabilities can be exploited directly from the VFMs. To this end, we propose the Transferable Video Attack (TVA), a temporal-aware adversarial attack method that leverages the temporal representation dynamics of VFMs to craft effective perturbations. TVA integrates a bidirectional contrastive learning mechanism to maximize the discrepancy between the clean and adversarial features, and introduces a temporal consistency loss that exploits motion cues to enhance the sequential impact of perturbations. TVA avoids the need to train expensive surrogate models or access to domain-specific data, thereby offering a more practical and efficient attack strategy. Extensive experiments across 24 video-related tasks demonstrate the efficacy of TVA against downstream models and MLLMs, revealing a previously underexplored security vulnerability in the deployment of video models.

Meng Shen, Minghao Wu, Deepu Rajan

Object hallucination is a significant challenge that hinders the application of large vision-language models (LVLMs) in practice. We hypothesize that one possible origin of hallucination is the model's tendency to prioritize text generation over meaningful interaction with images. To explore this, we examine the generation process and categorize text tokens into three groups: image-positive, invariant, and negative, based on their visual dependence on input image tokens. Our analysis reveals that most generated tokens are minimally influenced by the image information. This suggests that during the model's training stage, more emphasis is placed on learning how to follow textual instructions, rather than extracting information from images. Based on this finding, we propose adjusting the training weights of different tokens depending on their visual dependence to control hallucination. Additionally, we remove a portion of the training data that potentially contains more hallucinations as a data filtering strategy. Both methods achieve a reduction in hallucination without compromising response length or introducing additional computational costs during inference. We validate our methods across three LVLM variants, demonstrating the effectiveness and general applicability.

Ashutosh Anshul, Shreyas Gopal, Deepu Rajan et al.

Recent multimodal deepfake detection methods designed for generalization conjecture that single-stage supervised training struggles to generalize across unseen manipulations and datasets. However, such approaches that target generalization require pretraining over real samples. Additionally, these methods primarily focus on detecting audio-visual inconsistencies and may overlook intra-modal artifacts causing them to fail against manipulations that preserve audio-visual alignment. To address these limitations, we propose a single-stage training framework that enhances generalization by incorporating next-frame prediction for both uni-modal and cross-modal features. Additionally, we introduce a window-level attention mechanism to capture discrepancies between predicted and actual frames, enabling the model to detect local artifacts around every frame, which is crucial for accurately classifying fully manipulated videos and effectively localizing deepfake segments in partially spoofed samples. Our model, evaluated on multiple benchmark datasets, demonstrates strong generalization and precise temporal localization.

Chinthani Sugandhika, Chen Li, Deepu Rajan et al.

Spatio-temporal scene graph generation (ST-SGG) aims to model objects and their evolving relationships across video frames, enabling interpretable representations for downstream reasoning tasks such as video captioning and visual question answering. Despite recent advancements in DETR-style single-stage ST-SGG models, they still suffer from several key limitations. First, while these models rely on attention-based learnable queries as a core component, these learnable queries are semantically uninformed and instance-agnostically initialized. Second, these models rely exclusively on unimodal visual features for predicate classification. To address these challenges, we propose VOST-SGG, a VLM-aided one-stage ST-SGG framework that integrates the common sense reasoning capabilities of vision-language models (VLMs) into the ST-SGG pipeline. First, we introduce the dual-source query initialization strategy that disentangles what to attend to from where to attend, enabling semantically grounded what-where reasoning. Furthermore, we propose a multi-modal feature bank that fuses visual, textual, and spatial cues derived from VLMs for improved predicate classification. Extensive experiments on the Action Genome dataset demonstrate that our approach achieves state-of-the-art performance, validating the effectiveness of integrating VLM-aided semantic priors and multi-modal features for ST-SGG. We will release the code at https://github.com/LUNAProject22/VOST.

Chinthani Sugandhika, Chen Li, Deepu Rajan et al.

Large Video-Language Models (Video-LMs) have achieved impressive progress in multimodal understanding, yet their reasoning remains weakly grounded in space and time. We present Know-Show, a new benchmark designed to evaluate spatio-temporal grounded reasoning, the ability of a model to reason about actions and their semantics while simultaneously grounding its inferences in visual and temporal evidence. Know-Show unifies reasoning and localization within a single evaluation framework consisting of five complementary scenarios across spatial (person, object, person-object, and hand-object) and temporal dimensions. Built from Charades, Action Genome, and Ego4D with 2.5K human-authored questions, the benchmark exposes significant gaps between current Video-LMs and human reasoning. To bridge this gap, we propose GRAM, a training-free plug-in that augments Video-LMs with fine-grained grounding through attention-based video token selection and explicit timestamp encoding. Extensive experiments across open and closed Video-LMs (Qwen, VideoLLaVA, GPT-4o, and Gemini, etc.) reveal that existing models struggle to "show what they know" and vice versa, especially in fine-grained hand-object interactions. Know-Show establishes a unified standard for assessing grounded reasoning in video-language understanding and provides insights toward developing interpretable and reliable multimodal reasoning systems. We will release the dataset and the code at https://github.com/LUNAProject22/Know-Show.

Chen Li, Chinthani Sugandhika, Yeo Keat Ee et al.

Existing human motion Q\&A methods rely on explicit program execution, where the requirement for manually defined functional modules may limit the scalability and adaptability. To overcome this, we propose an implicit program-guided motion reasoning (IMoRe) framework that unifies reasoning across multiple query types without manually designed modules. Unlike existing implicit reasoning approaches that infer reasoning operations from question words, our model directly conditions on structured program functions, ensuring a more precise execution of reasoning steps. Additionally, we introduce a program-guided reading mechanism, which dynamically selects multi-level motion representations from a pretrained motion Vision Transformer (ViT), capturing both high-level semantics and fine-grained motion cues. The reasoning module iteratively refines memory representations, leveraging structured program functions to extract relevant information for different query types. Our model achieves state-of-the-art performance on Babel-QA and generalizes to a newly constructed motion Q\&A dataset based on HuMMan, demonstrating its adaptability across different motion reasoning datasets. Code and dataset are available at: https://github.com/LUNAProject22/IMoRe.

Heqing Zou, Fengmao Lv, Desheng Zheng et al.

Multilingual speech emotion recognition aims to estimate a speaker's emotional state using a contactless method across different languages. However, variability in voice characteristics and linguistic diversity poses significant challenges for zero-shot speech emotion recognition, especially with multilingual datasets. In this paper, we propose leveraging contrastive learning to refine multilingual speech features and extend large language models for zero-shot multilingual speech emotion estimation. Specifically, we employ a novel two-stage training framework to align speech signals with linguistic features in the emotional space, capturing both emotion-aware and language-agnostic speech representations. To advance research in this field, we introduce a large-scale synthetic multilingual speech emotion dataset, M5SER. Our experiments demonstrate the effectiveness of the proposed method in both speech emotion recognition and zero-shot multilingual speech emotion recognition, including previously unseen datasets and languages.

Meng Shen, Yake Wei, Jianxiong Yin et al.

Training multimodal models requires a large amount of labeled data. Active learning (AL) aim to reduce labeling costs. Most AL methods employ warm-start approaches, which rely on sufficient labeled data to train a well-calibrated model that can assess the uncertainty and diversity of unlabeled data. However, when assembling a dataset, labeled data are often scarce initially, leading to a cold-start problem. Additionally, most AL methods seldom address multimodal data, highlighting a research gap in this field. Our research addresses these issues by developing a two-stage method for Multi-Modal Cold-Start Active Learning (MMCSAL). Firstly, we observe the modality gap, a significant distance between the centroids of representations from different modalities, when only using cross-modal pairing information as self-supervision signals. This modality gap affects data selection process, as we calculate both uni-modal and cross-modal distances. To address this, we introduce uni-modal prototypes to bridge the modality gap. Secondly, conventional AL methods often falter in multimodal scenarios where alignment between modalities is overlooked. Therefore, we propose enhancing cross-modal alignment through regularization, thereby improving the quality of selected multimodal data pairs in AL. Finally, our experiments demonstrate MMCSAL's efficacy in selecting multimodal data pairs across three multimodal datasets.

Chinthani Sugandhika, Chen Li, Deepu Rajan et al.

Graph based representation has been widely used in modelling spatio-temporal relationships in video understanding. Although effective, existing graph-based approaches focus on capturing the human-object relationships while ignoring fine-grained semantic properties of the action components. These semantic properties are crucial for understanding the current situation, such as where does the action takes place, what tools are used and functional properties of the objects. In this work, we propose a graph-based representation called Situational Scene Graph (SSG) to encode both human-object relationships and the corresponding semantic properties. The semantic details are represented as predefined roles and values inspired by situation frame, which is originally designed to represent a single action. Based on our proposed representation, we introduce the task of situational scene graph generation and propose a multi-stage pipeline Interactive and Complementary Network (InComNet) to address the task. Given that the existing datasets are not applicable to the task, we further introduce a SSG dataset whose annotations consist of semantic role-value frames for human, objects and verb predicates of human-object relations. Finally, we demonstrate the effectiveness of our proposed SSG representation by testing on different downstream tasks. Experimental results show that the unified representation can not only benefit predicate classification and semantic role-value classification, but also benefit reasoning tasks on human-centric situation understanding. We will release the code and the dataset soon.

Hui Lu, Yi Yu, Shijian Lu et al.

Temporal Action Detection (TAD) aims to identify and localize actions by determining their starting and ending frames within untrimmed videos. Recent Structured State-Space Models such as Mamba have demonstrated potential in TAD due to their long-range modeling capability and linear computational complexity. On the other hand, structured state-space models often face two key challenges in TAD, namely, decay of temporal context due to recursive processing and self-element conflict during global visual context modeling, which become more severe while handling long-span action instances. Additionally, traditional methods for TAD struggle with detecting long-span action instances due to a lack of global awareness and inefficient detection heads. This paper presents MambaTAD, a new state-space TAD model that introduces long-range modeling and global feature detection capabilities for accurate temporal action detection. MambaTAD comprises two novel designs that complement each other with superior TAD performance. First, it introduces a Diagonal-Masked Bidirectional State-Space (DMBSS) module which effectively facilitates global feature fusion and temporal action detection. Second, it introduces a global feature fusion head that refines the detection progressively with multi-granularity features and global awareness. In addition, MambaTAD tackles TAD in an end-to-end one-stage manner using a new state-space temporal adapter(SSTA) which reduces network parameters and computation cost with linear complexity. Extensive experiments show that MambaTAD achieves superior TAD performance consistently across multiple public benchmarks.

Heqing Zou, Meng Shen, Chen Chen et al.

Multimodal learning aims to imitate human beings to acquire complementary information from multiple modalities for various downstream tasks. However, traditional aggregation-based multimodal fusion methods ignore the inter-modality relationship, treat each modality equally, suffer sensor noise, and thus reduce multimodal learning performance. In this work, we propose a novel multimodal contrastive method to explore more reliable multimodal representations under the weak supervision of unimodal predicting. Specifically, we first capture task-related unimodal representations and the unimodal predictions from the introduced unimodal predicting task. Then the unimodal representations are aligned with the more effective one by the designed multimodal contrastive method under the supervision of the unimodal predictions. Experimental results with fused features on two image-text classification benchmarks UPMC-Food-101 and N24News show that our proposed Unimodality-Supervised MultiModal Contrastive UniS-MMC learning method outperforms current state-of-the-art multimodal methods. The detailed ablation study and analysis further demonstrate the advantage of our proposed method.

Dilip K. Prasad, Huixu Dong, Deepu Rajan et al.

Maritime vessels equipped with visible and infrared cameras can complement other conventional sensors for object detection. However, application of computer vision techniques in maritime domain received attention only recently. The maritime environment offers its own unique requirements and challenges. Assessment of the quality of detections is a fundamental need in computer vision. However, the conventional assessment metrics suitable for usual object detection are deficient in the maritime setting. Thus, a large body of related work in computer vision appears inapplicable to the maritime setting at the first sight. We discuss the problem of defining assessment metrics suitable for maritime computer vision. We consider new bottom edge proximity metrics as assessment metrics for maritime computer vision. These metrics indicate that existing computer vision approaches are indeed promising for maritime computer vision and can play a foundational role in the emerging field of maritime computer vision.

Raj Kumar Gupta, Alex Yong-Sang Chia, Deepu Rajan et al.

In this paper, we present a color transfer algorithm to colorize a broad range of gray images without any user intervention. The algorithm uses a machine learning-based approach to automatically colorize grayscale images. The algorithm uses the superpixel representation of the reference color images to learn the relationship between different image features and their corresponding color values. We use this learned information to predict the color value of each grayscale image superpixel. As compared to processing individual image pixels, our use of superpixels helps us to achieve a much higher degree of spatial consistency as well as speeds up the colorization process. The predicted color values of the gray-scale image superpixels are used to provide a 'micro-scribble' at the centroid of the superpixels. These color scribbles are refined by using a voting based approach. To generate the final colorization result, we use an optimization-based approach to smoothly spread the color scribble across all pixels within a superpixel. Experimental results on a broad range of images and the comparison with existing state-of-the-art colorization methods demonstrate the greater effectiveness of the proposed algorithm.

Jubin Johnson, Hisham Cholakkal, Deepu Rajan

Sampling-based alpha matting methods have traditionally followed the compositing equation to estimate the alpha value at a pixel from a pair of foreground (F) and background (B) samples. The (F,B) pair that produces the least reconstruction error is selected, followed by alpha estimation. The significance of that residual error has been left unexamined. In this letter, we propose a video matting algorithm that uses L1-regularized reconstruction error of F and B samples as a measure of the alpha matte. A multi-frame non-local means framework using coherency sensitive hashing is utilized to ensure temporal coherency in the video mattes. Qualitative and quantitative evaluations on a dataset exclusively for video matting demonstrate the effectiveness of the proposed matting algorithm.

Hisham Cholakkal, Jubin Johnson, Deepu Rajan

Top-down saliency models produce a probability map that peaks at target locations specified by a task/goal such as object detection. They are usually trained in a fully supervised setting involving pixel-level annotations of objects. We propose a weakly supervised top-down saliency framework using only binary labels that indicate the presence/absence of an object in an image. First, the probabilistic contribution of each image region to the confidence of a CNN-based image classifier is computed through a backtracking strategy to produce top-down saliency. From a set of saliency maps of an image produced by fast bottom-up saliency approaches, we select the best saliency map suitable for the top-down task. The selected bottom-up saliency map is combined with the top-down saliency map. Features having high combined saliency are used to train a linear SVM classifier to estimate feature saliency. This is integrated with combined saliency and further refined through a multi-scale superpixel-averaging of saliency map. We evaluate the performance of the proposed weakly supervised topdown saliency and achieve comparable performance with fully supervised approaches. Experiments are carried out on seven challenging datasets and quantitative results are compared with 40 closely related approaches across 4 different applications.

Hisham Cholakkal, Jubin Johnson, Deepu Rajan

We propose a framework for top-down salient object detection that incorporates a tightly coupled image classification module. The classifier is trained on novel category-aware sparse codes computed on object dictionaries used for saliency modeling. A misclassification indicates that the corresponding saliency model is inaccurate. Hence, the classifier selects images for which the saliency models need to be updated. The category-aware sparse coding produces better image classification accuracy as compared to conventional sparse coding with a reduced computational complexity. A saliency-weighted max-pooling is proposed to improve image classification, which is further used to refine the saliency maps. Experimental results on Graz-02 and PASCAL VOC-07 datasets demonstrate the effectiveness of salient object detection. Although the role of the classifier is to support salient object detection, we evaluate its performance in image classification and also illustrate the utility of thresholded saliency maps for image segmentation.

Jubin Johnson, Ehsan Shahrian Varnousfaderani, Hisham Cholakkal et al.

Existing color sampling based alpha matting methods use the compositing equation to estimate alpha at a pixel from pairs of foreground (F) and background (B) samples. The quality of the matte depends on the selected (F,B) pairs. In this paper, the matting problem is reinterpreted as a sparse coding of pixel features, wherein the sum of the codes gives the estimate of the alpha matte from a set of unpaired F and B samples. A non-parametric probabilistic segmentation provides a certainty measure on the pixel belonging to foreground or background, based on which a dictionary is formed for use in sparse coding. By removing the restriction to conform to (F,B) pairs, this method allows for better alpha estimation from multiple F and B samples. The same framework is extended to videos, where the requirement of temporal coherence is handled effectively. Here, the dictionary is formed by samples from multiple frames. A multi-frame graph model, as opposed to a single image as for image matting, is proposed that can be solved efficiently in closed form. Quantitative and qualitative evaluations on a benchmark dataset are provided to show that the proposed method outperforms current state-of-the-art in image and video matting.