Temporal sentence grounding in videos (TSGV) is a challenging task that aims to match text queries with semantically relevant segments in untrimmed videos. However, existing methods face limitations in modeling modality features, which constrains the expressive power of candidate moment features. To address this challenge, we propose a novel Enhanced Feature Interaction Network (EFIN) that effectively captures semantic information within each modality and aligns relationships between modalities. Additionally, EFIN enhances the fusion of information between candidate moments and modality features. Specifically, our model begins by extracting modality features to generate candidate moments as priors. Building upon these modality features, we introduce an enhanced feature encoder to extract semantic information within each modality, thereby improving intra-modality feature representation. Simultaneously, the encoder captures alignment relationships between modalities to optimize cross-modality feature representation, enhancing the overall modeling capacity of modality features. Moreover, we design an information fusion module to enrich the comprehension of modality information for candidate moments. Extensive experiments on four benchmark datasets demonstrate the superiority of our proposed EFIN model. Notably, EFIN achieves a maximum performance improvement of approximately 1.67% and 1.91% across different evaluation metrics on TACoS dataset.

In recent years, many excellent text classification models have been proposed by researchers to solve natural language processing tasks. In texts, semantic information at different levels and granularities constitutes the multi-dimensional semantic information of the text, which determines its classification effect. However, existing text classification models have several limitations. Firstly, they cannot extract the multi-dimensional semantic information from the text. Secondly, they cannot determine the weight of the multi-dimensional semantic information on the text's semantic information. To address these limitations, this paper proposes a new deep learning neural network model called the multi-dimensional semantic-based text classification model (MSTCM), which comprises an embedding module, a semantic processing module, and a prediction module. After processing the text through the embedding module, it passes through several sub-modules and layers in the semantic processing module to extract the multi-dimensional semantic information of the text and its weight on the text's semantic information using the attention mechanism. This information is used to generate a text vector, which is then utilized by the prediction module for text classification. To verify the effectiveness of MSTCM, experiments were conducted on several text classification tasks and compared with various text classification methods. The experimental results demonstrate that MSTCM performs better than other methods on most tasks.