Benchmark Radar
AI BENCHMARK PROFILE

TSC-Bench

General AIMultimodal Perception

TSC-Bench is a benchmark for triple-shot composition, generating establishing, medium, and close-up crops from human-centric images with shot descriptions. It contains 1.2k expert-annotated test cases.

Released
2026-06-04
Readiness
Paper only
Primary field
General AI

Why it matters

Multi-shot composition is valuable for creative workflows, but existing benchmarks focus on single crops. TSC-Bench could evaluate narrative-driven cropping capabilities.

Motivation

Prior work on aesthetic composition typically produces a single aesthetically pleasing crop, overlooking the narrative value of composing multiple shots from one scene.

Primary resources

Benchmark Radar records only publicly supported details and links back to primary sources for verification.