PixLore: A Dataset-driven Approach to Rich Image Captioning

Abstract

In the domain of vision-language integration, generating detailed image captions poses a significant challenge due to the lack of curated and rich datasets. This study introduces PixLore, a novel method that leverages Querying Transformers through the fine-tuning of the BLIP-2 model using the LoRa method on a standard commercial GPU. The followed approach, which involves training on a carefully assembled dataset from state-of-the-art Computer Vision models combined and augmented by ChatGPT, addresses the question of whether intricate image understanding can be achieved with an ensemble of smaller-scale models, referred to as Knowledge Stitching. Comparative evaluations against major models such as GPT-4 and Google Bard demonstrate that PixLore-2.7B, despite having considerably fewer parameters, is rated higher than the existing State-of-the-Art models in over half of the assessments. Precisely, PixLore outperform Bard and BLIP-2, which score approximately 35.18% and 27.98% lower than PixLore in the task of image captioning. This research not only presents a groundbreaking approach but also highlights the importance of well-curated datasets in enhancing the performance of smaller models.