Text-prompted methods have been proven effective in enhancing the performance of medical image segmentation tasks. However, text-prompted segmentation of pulmonary infection regions still faces significant challenges in practical applications: the high cost of comprehensive medical prompts leads to frequent occurrences of missing medical text modalities in clinical testing scenarios. To address this issue, we propose a novel Knowledge Proxy Learning with Missing Modalities for Medical Image Segmentation (MMKPL-Seg). This model employs a learnable Memory Bank to automatically extract and store textual information during the training phase. Then, we introduce the Knowledge Proxy Learning (KPL) for both complete and missing modality scenarios, it retrieves relevant knowledge from the Memory Bank and integrates it into the encoder. By iteratively training on both single-image and image-text paired samples, the model enables single-image inputs to retain potential semantic information during the testing phase. Extensive evaluations on the QaTa-COV19 and MosMedData+ datasets demonstrate that our model achieves state-of-the-art performance compared to both unimodal and previous multimodal methods. Notably, MMKPL-Seg maintains stable segmentation performance even when text modalities are completely absent, making it highly adaptable to the common occurrence of incomplete textual information in clinical practice. This characteristic underscores its substantial practical value.

In recent years, remarkable progress has been made in remote sensing image-text retrieval (RSITR), which has transitioned from relying on compact global features to more fine-grained local features representing salient objects in images. However, existing methods typically concentrate only on significant entity information in remote sensing images as local features, overlooking the correlations between entities, resulting in isolated entity information. Moreover, there is ample room for exploring text local features.To address these issues, this paper presents an Entity Spatial Relation enhancement Network (ESRN), leveraging global and local entity features in remote sensing images and texts. For local feature processing, Graph Convolutional Network (GCN) is employed to aggregate the correlation between entity features and entity information in remote sensing images, enhancing entity information learning. The self-attention mechanism is used to model the remote dependency of entity keywords and spatial orientation semantic relations in text, strengthening the text representation ability. A strategy of proportionally adding different levels of features is proposed to enhance the representation of salient features and reduce noise interference.The approach was evaluated on two renowned remote sensing datasets, RSICD and RSITMD, validating the model's capacity to perceive the semantics of remote sensing images and text entities. Performance comparison, ablation experiments, and visualization analysis convincingly demonstrate the state-of-the-art performance of the ESRN method in the RSITR task.