This research focuses firstly on the improvement of Image Classification using Large Language Models, with the goal to secondly, incorporate findings into Open-Vocabulary Object Detection. The research focuses on the following areas:

1. Zero-Shot Image Classification Accuracy Improvements from Large Language Model generated Side Information.
2. Few-Shot CLIP (Contrastive Language-Image Pre-training) Adaptation for Image Classification.
3. Few-Shot, Fine-Grain, Spatial Alignment-Based CLIP Adaptation.
4. Transfer of CLIP Adaptation Improvements in Image Classification to Open-Vocabulary Object Detection.