This study introduces a multimodal large language model capable of not only accomplishing various heart sound tasks but also providing reasoning, marking an advancement in the field of medical diagnostics. The model’s innovation stems from a collaboration with experts to collect a novel dataset designed specifically for reasoning tasks, addressing the limitations of existing datasets that lacked this capability. Our model integrates multiple novel methodologies to enhance diagnostic accuracy, including the incorporation of knowledge from relevant textbooks through pre-training, the employment of an audio feature extractor optimized for heart sound-text alignment, and a logit adjustment loss tailored for large language model to mitigate the challenge of imbalanced data categories. This approach not only sets a new standard for heart sound analysis but also paves the way for more interpretable and comprehensive diagnostic models in healthcare.

The weakly-supervised audio-visual video parsing (AVVP) task aims toparse a video into temporal events and predict their modality-specific categories. Current works primarily focus on refining training strategies and follow the framework fusing signals only at the segment level. However, they miss the point that video events, being composed of consecutive segments, require the integration of both local and global contexts to fully capture their essence. In this letter, we present theLocal-GlobalFusionNetwork (LGFNet), designed to facilitate multi-level interaction between audio and visual signals. Specifically, we create a two-dimensional map to generate multi-scale event proposals for both audio and visual modalities. Subsequently, we fuse audio and visual signals at both segment and event levels with a novel boundary-aware feature aggregation method, enabling the simultaneous capture of local and global information. To enhance the temporal alignment between the two modalities, we employ segment-level and event-level contrastive learning. In-depth experiments demonstrate the superiority of our LGFNet.

Current mainstream vision-language (VL) tracking framework consists of three parts,i.e., a visual feature extractor, a language feature extractor, and a fusion model. To pursue better performance, a natural modus operandi for VL tracking is employing customized and heavier unimodal encoders, and multi-modal fusion models. Albeit effective, existing VL trackers separate feature extraction and feature integration, resulting in extracted features that lack semantic guidance and have limited target-aware capability in complex scenarios, e.g., similar distractors and extreme illumination. In this work, inspired by the recent success of exploring foundation models with unified architecture for both natural language and computer vision tasks, we propose an All-in-One framework, which learns joint feature extraction and interaction by adopting a unified transformer backbone. Specifically, we mix raw vision and language signals to generate language-injected vision tokens, which we then concatenate before feeding into the unified backbone architecture. This approach achieves feature integration in a unified backbone, removing the need for carefully-designed fusion modules and resulting in a more effective and efficient VL tracking framework. To further improve the learning efficiency, we introduce a multi-modal alignment module based on cross-modal and intra-modal contrastive objectives, providing more reasonable representations for the unified All-in-One transformer backbone. Extensive experiments on five benchmarks, i.e., OTB99-L, TNL2K, LaSOT, LaSOTExt and WebUAV-3M, demonstrate the superiority of the proposed tracker against existing state-of-the-art (SOTA) methods on VL tracking. Codes will be available at https://github.com/983632847/All-in-One here.

Video moment retrieval aims to retrieve a target moment from an untrimmed video that semantically corresponds to the given language query. Existing methods commonly treat it as a regression task or a ranking task from the perspective of computer vision. Most of these works neglect comprehensive relations between video content and language context at a multi-granularity level and fail to efficiently model temporal relations among different video moments. In this paper, we formulate video moment retrieval into video reading comprehension by treating the input video as a text passage and language query as a question. To tackle the above impediments, we propose a Comprehensive Relation-aware Network (CRNet) to perceive comprehensive relations from extensive aspects. Specifically, we unite visual and textual features simultaneously at both clip-level and moment-level to thoroughly exploit inter-modality information, leading to a coarse-and-fine cross-modal interaction. Moreover, a background suppression module is introduced to restrain irrelevant background clips, meanwhile, a novel IoU attention mechanism and graph attention layer are efficiently devised to focus on the dependencies among highly-correlated video moments for the best choice selection. In-depth experiments on three public datasets TACoS, ActivityNet Captions, and Charades-STA demonstrate the superiority of our solution.

Moment retrieval in videos is a challenging task that aims to retrieve the most relevant video moment in an untrimmed video given a sentence description. Previous methods tend to perform self-modal learning and cross-modal interaction in a coarse manner, which neglect fine-grained clues contained in video content, query context, and their alignment. To this end, we propose a novel Multi-Granularity Perception Network (MGPN) that perceives intra-modality and inter-modality information at a multi-granularity level. Specifically, we formulate moment retrieval as a multi-choice reading comprehension task and integrate human reading strategies into our framework. A coarse-grained feature encoder and a co-attention mechanism are utilized to obtain a preliminary perception of intra-modality and inter-modality information. Then a fine-grained feature encoder and a conditioned interaction module are introduced to enhance the initial perception inspired by how humans address reading comprehension problems. Moreover, to alleviate the huge computation burden of some existing methods, we further design an efficient choice comparison module and reduce the hidden size with imperceptible quality loss. Extensive experiments on Charades-STA, TACoS, and ActivityNet Captions datasets demonstrate that our solution outperforms existing state-of-the-art methods.

Video grounding aims at localizing the temporal moment related to the given language description, which is very helpful to many cross-modal content understanding applications like visual question answering and sentence-video search. Existing approaches usually directly regress the temporal boundaries of an event described by a query sentence in the video sequence. This direct regression manner often encounters a large decision space due to diverse target events and variable video durations, leading to inaccurate localization as well as inefficient grounding. This paper presents an efficient framework termed from which to where to facilitate video grounding. The core idea is imitating the reading comprehension process to gradually narrow the decision space, in what we decompose the direct regression into two steps. The “which” step first roughly selects a candidate area by evaluating which video segment in the predefined set is closest to the ground truth. To this end, we formulate this step into a multi-choice reading comprehension problem and propose a criterion to select the best-matched segment. In this way, the excessive decision space is effectively reduced. The “where” step aims to precisely regress the temporal boundary of the selected video segment from the shrunk decision space. We thus introduce a triple-span representation for each candidate video segment to use the regional context for better boundary regression. The “which” and “where” steps can be combined into a unified framework and learned end-to-end, leading to an efficient video grounding system. Extensive experiments on Charades-STA, ActivityNet-Captions, and TACoS benchmarks clearly demonstrate the effectiveness of our framework.

Temporal language grounding in videos aims to localize the temporal span relevant to the given query sentence.Previous methods treat it either as a boundary regression task or a span extraction task.This paper will formulate temporal language grounding into video reading comprehension and propose a Relation-aware Network (RaNet) to address it.This framework aims to select a video moment choice from the predefined answer set with the aid of coarse-and-fine choice-query interaction and choice-choice relation construction.A choicequery interactor is proposed to match the visual and textual information simultaneously in sentence-moment and token-moment levels, leading to a coarse-and-fine cross-modal interaction.Moreover, a novel multi-choice relation constructor is introduced by leveraging graph convolution to capture the dependencies among video moment choices for the best choice selection.Extensive experiments on ActivityNet-Captions, TACoS, and Charades-STA demonstrate the effectiveness of our solution.Codes will be available at https: //github.com/Huntersxsx/RaNet.