Binary Verification for Zero-Shot Vision

TL;DR

This paper introduces a training-free binary verification workflow for zero-shot vision tasks using off-the-shelf vision-language models, improving performance across multiple tasks without additional training.

Contribution

The authors propose a novel, training-free workflow combining quantization and binarization steps to enhance zero-shot vision performance with existing models.

Findings

Quantization to MCQ significantly improves task accuracy.

True/False binarization further boosts performance.

The workflow generalizes across various vision tasks.

Abstract

We propose a training-free, binary verification workflow for zero-shot vision with off-the-shelf VLMs. It comprises two steps: (i) quantization, which turns the open-ended query into a multiple-choice question (MCQ) with a small, explicit list of unambiguous candidates; and (ii) binarization, which asks one True/False question per candidate and resolves deterministically: if exactly one is True, select it; otherwise, revert to an MCQ over the remaining plausible candidates. We evaluate the workflow on referring expression grounding (REC), spatial reasoning (Spatial-Map, Spatial-Grid, Spatial-Maze), and BLINK-Jigsaw. Relative to answering open-ended queries directly, quantization to MCQ yields large gains, and True/False binarization provides a consistent additional boost. Across all tasks, the same workflow produces significant improvements, indicating generality. We further integrate the proposed REC workflow into a real-world video processing and editing system, and present the system architecture and end-to-end pipeline in the paper. Together, these components yield a simple and unified workflow that emphasizes inference-time design over task-specific training. It offers a practical, drop-in path to stronger zero-shot vision with today's VLMs.