ResearchBenchmarksObject Selection on Real Scenes sparse clutter (2-4 objects)Object Selection on Real Scenes sparse clutter (2-4 objects)Follow65AccuracyCLIP-Grounding3945.7552.559.25Mar 3, 2026Evaluation resultsMethodMethodLinksAccuracyCLIP-Grounding2026.0365SRE2026.0360GPT-4o2026.0340