Multimodal Contrastive Learning for Remote Sensing Image Feature Extraction Based on Relaxed Positive Samples.

Traditional multimodal contrastive learning brings text and its corresponding image closer together as a positive pair, where the text typically consists of fixed sentence structures or specific descriptive statements, and the image features are generally global features (with some fine-grained work using local features). Similar to unimodal self-supervised contrastive learning, this approach can be seen as enforcing a strict identity constraint in a multimodal context. However, due to the inherent complexity of remote sensing images, which cannot be easily described in a single sentence, and the fact that remote sensing images contain rich ancillary information beyond just object features, this strict identity constraint may be insufficient.

Parameters optimization for decontamination and fine physical regeneration pathways of polypropylene plastics from waste lunchboxes.

Due to the development of the food delivery industry, a large amount of waste lunchboxes made of homo polypropylene (PP) plastic have been generated. This study developed a new technological strategy to effectively regenerate PP from waste lunchboxes. Through response surface curve analysis, it was found that under the optimal process conditions of hot alkali washing at 80 ℃, 30 min, and pH 13, the optimal contact angle was 65.