Article

CAT: Content-Adaptive Image Tokenization

2025-01-17

Abstract excerpt

Most existing image tokenizers encode images into a fixed number of tokens or patches, overlooking the inherent variability in image complexity. To address this, we introduce Content-Adaptive Tokenizer (CAT), which dynamically adjusts representation capacity based on the image content and encodes simpler images into fewer tokens. We design a caption-based evaluation system that leverages large language models (LLM...

Topics

Open a Topic to create a Post that cites this publication.

Radiomics and Machine Learning in Medical Imaging

Identifiers and source

Literature Corpus work: 804f7323-86a7-5915-bd05-610ef82867fd
DOI: 10.32388/wcbnq2

Open publication

Related research

Semantic proximity does not establish scientific evidence.

Click a neighbor to travelStep 1 · 12 closest

CAT: Content-Adaptive Image TokenizationDOI 10.32388/wcbnq2Read

LinGen-Uni: A Universal Linear-Complexity Framework for High-Resolution Minute-Length Text-to-Video GenerationSemantic proximity Read

Dynamic Token Clustering: A Novel Architectural Enhancement for Large Language ModelsSemantic proximity Read

MiniCausal-T2V: Towards Ultra-Low Latency and Memory-Efficient Causal Video Generation on Edge DevicesSemantic proximity Read

HEADHUNTER: Training-Free Annotated Dataset Synthesis via Self-Guided Diffusion Transformer Attention Head SelectionSemantic proximity Read

HEADHUNTER: Training-Free Annotated Dataset Synthesis via Self-Guided Diffusion Transformer Attention Head SelectionSemantic proximity Read

DiscoverNet: Unsupervised Video Object Discovery through Discrete Visual Tokenization and Language Model-Inspired ReconstructionSemantic proximity Read

Rethinking Convolutional Semantics for Image Caption Generation Beyond Recurrent ParadigmsSemantic proximity Read

Adaptive Semantic Fusion for Contextual Image CaptioningSemantic proximity Read

Tarsier2: Advancing Large Vision-Language Models from Detailed Video Descriptions to Comprehensive Video UnderstandingSemantic proximity Read

Enhancing Caption Fidelity via Explanation-Guided Captioning with Vision-Language Fine-TuningSemantic proximity Read

A Comparative Survey of CNN-LSTM Architectures for Image CaptioningSemantic proximity Read

An Efficient and Training-Free Approach for Subject-Driven Text-to-Image GenerationSemantic proximity Read