Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Slot-MLLM: Object-Centric Visual Tokenization for Multimodal LLM

About

Recently, multimodal large language models (MLLMs) have emerged as a key approach in achieving artificial general intelligence. In particular, vision-language MLLMs have been developed to generate not only text but also visual outputs from multimodal inputs. This advancement requires efficient image tokens that LLMs can process effectively both in input and output. However, existing image tokenization methods for MLLMs typically capture only global abstract concepts or uniformly segmented image patches, restricting MLLMs' capability to effectively understand or generate detailed visual content, particularly at the object level. To address this limitation, we propose an object-centric visual tokenizer based on Slot Attention specifically for MLLMs. In particular, based on the Q-Former encoder, diffusion decoder, and residual vector quantization, our proposed discretized slot tokens can encode local visual details while maintaining high-level semantics, and also align with textual data to be integrated seamlessly within a unified next-token prediction framework of LLMs. The resulting Slot-MLLM demonstrates significant performance improvements over baselines with previous visual tokenizers across various vision-language tasks that entail local detailed comprehension and generation. Notably, this work is the first demonstration of the feasibility of object-centric slot attention performed with MLLMs and in-the-wild natural images.

Donghwan Chi, Hyomin Kim, Yoonjin Oh, Yongjin Kim, Donghoon Lee, Daejin Jo, Jongmin Kim, Junyeob Baek, Sungjin Ahn, Sungwoong Kim• 2025

Related benchmarks

TaskDatasetResultRank
Object Hallucination EvaluationPOPE--
2019
Multimodal UnderstandingMMBench--
847
Text-to-Image GenerationGenEval
GenEval Score57
442
Visual Question AnsweringGQA (test)
Accuracy58.8
197
Multimodal UnderstandingSEED-Bench Image--
143
Multimodal UnderstandingSEEDBench2 Plus
Accuracy34.4
138
Multimodal UnderstandingPOPE
POPE Score0.83
112
Multimodal UnderstandingMMMU
MMMU Score46.7
102
Multimodal UnderstandingMME Perception
MME-P Score1.47e+3
59
Multi-modal Vision-Language UnderstandingGQA
Accuracy58.8
51
Showing 10 of 19 rows

Other info

Follow for update