Research map / Detection and segmentation / Open-set and few-shot detection
Ins-DetCLIP: Aligning Detection Model to Follow Human-Language Instruction
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Introduces Instruction-oriented Object Detection (IOD), a task where detectors follow natural-language instructions to locate relevant objects, and proposes Ins-DetCLIP to align a detection model to such instructions.
Problem
Traditional open-vocabulary object detection requires users to supply explicit category name lists, whereas real human-computer interaction needs detectors that understand natural-language instructions and contextual reasoning to locate relevant objects.
Approach
Defines the instruction-oriented object detection task and aligns a CLIP-based detection model to follow human-language instructions, going beyond category-name prompts to instruction comprehension and contextual reasoning.
Experimental setup
Builds on: DetCLIP
Closest papers by meaning
- InstructDET: Diversifying Referring Object Detection with Generalized Instructions (ICLR 2024 · similarity 0.90)
- LINK: Learning Instance-level Knowledge from Vision-Language Models for Human-Object Interaction Detection (ICLR 2026 · similarity 0.90)
- DetCLIPv3: Towards Versatile Generative Open-vocabulary Object Detection (CVPR 2024 · similarity 0.89)
- Open-Det: An Efficient Learning Framework for Open-Ended Detection (ICML 2025 · similarity 0.88)
- DetKDS: Knowledge Distillation Search for Object Detectors (ICML 2024 · similarity 0.87)
- Generative Region-Language Pretraining for Open-Ended Object Detection (CVPR 2024 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.