UMass-Rescue / ImageSearch_CLIP

A system that creates clip embedding vectors for a large corpus of images, enabling efficient retrieval of images based on natural language text.
0 stars 1 forks source link

Enhance Text-to-Image Search with Captioned Data #63

Open sahithi-66 opened 2 weeks ago

sahithi-66 commented 2 weeks ago

Description: Integrate captioned datasets into the image retrieval system to enhance text-to-image search capabilities. The task focuses on improving the semantic alignment between text queries and image content by leveraging rich contextual information from captions. The goal is to fine-tune the model on captioned data to enable more accurate and contextually relevant image retrieval for text-based queries.

Outcome: Integrate and fine-tune the model using captioned datasets to improve semantic alignment in text-to-image searches. Enhance the model's ability to process complex text queries and return relevant images based on the context provided in captions. Evaluate the impact of captioned data on the relevance and accuracy of text-based image retrieval results. Document improvements in retrieval performance and assess the effectiveness of captioned data in handling more nuanced queries.