University of Surrey Develops Sketch-Based Object Detection Tool for Machine Learning.

Researchers from the University of Surrey have developed a framework for detecting objects based on sketches, allowing for specific object identification in a scene. The model operates in a zero-shot fashion, without requiring extra boundary boxes or class labels. The researchers combined CLIP and sketch-based image retrieval (SBIR) to create a sketch-aware detector that outperforms supervised and weakly supervised object detectors in a zero-shot setting. The framework is instance-aware and part-aware, and bridges the gap between object and image levels.