Fast Zero Shot Object Detection with OpenAI CLIP

This course demonstrates how to use OpenAI CLIP for zero-shot image localization and object detection. It covers preprocessing, model initialization, sliding-window scoring, prompt changes, and bounding-box visualization in Python without fine-tuning.

Syllabus

Early Progress in Computer Vision
Classification vs. Localization and Detection
Zero Shot with OpenAI CLIP
Zero Shot Object Localization with OpenAI CLIP
Localization with Occlusion Algorithm
Zero Shot Object Detection with OpenAI CLIP
Data Preprocessing for CLIP
Initializing OpenAI CLIP in Python
Clipping the Localization Visual
Applying Scores for Visual
Object Localization with New Prompt
Zero Shot Object Detection in Python
Creating Bounding Boxes with Matplotlib
Object Detection Code
Object Detection Results
Trends in Multi-Modal ML