Pocket-Dentist: On-Device Dental Image Understanding via Efficient Multimodal Large Language Models
Quick Take
Pocket-Dentist introduces an efficient benchmark for dental multimodal question answering, demonstrating that compact VLMs like the 2B-parameter model outperform larger models in accuracy and computational cost. Deployed on an iPhone 17 Pro, it processes samples in 4.31 seconds, reducing latency by 4.9-fold and memory usage by 2.3-fold compared to a 7B baseline.
Key Points
- Pocket-Dentist benchmarks three datasets with 1,159 patients across five task types.
- Compact VLMs outperform larger models in dental image understanding accuracy.
- Finetuned Pocket-Dentist-2B processes samples in 4.31 seconds on iPhone 17 Pro.
- Latency reduced by 4.9-fold and memory use by 2.3-fold compared to a 7B model.
- Efficient deployment supports timely dental screening in non-specialist settings.
Article Excerpt
From source RSS / original summaryarXiv:2605. 29299v1 Announce Type: new Abstract: Evaluations of dental vision-language models remain fragmented across datasets, task definitions and metrics, and often ignore their computational cost. This limits their widespread deployment for dental screening outside specialist centres, where timely inference, limited hardware, and local handling of patient images are vital for practical, privacy-preserving clinical prescreening.
Here we present Pocket-Dentist, an efficiency-aware benchmark for dental multimodal question answering that brings together three datasets spanning approximately 1,159 patients, five task types and seven metrics. Across typical 14 VLMs, our results reveals an interesting observation: compact VLMs (e. g. , 2B-parameter models) outperform larger VLMs in accuracy while requiring substantially lower computational costs in dental image understanding.
Deployed locally on an iPhone 17 Pro, our finetuned compact VLM Pocket-Dentist-2B processed each sample in 4. 31 s, reducing latency by 4. 9-fold and memory use by 2. 3-fold compared with a 7B baseline.
Reader Mode unavailable (could not extract clean content).
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from arXiv cs.CV
See more →Evi-Steer: Learning to Steer Biomedical Vision-Language Models through Efficient and Generalizable Evidential Tuning
Evi-Steer introduces a novel evidential tuning framework for BiomedCLIP, achieving 0.11% parameter updates while enhancing uncertainty-aware fine-tuning. It outperforms state-of-the-art methods across 15 biomedical imaging datasets, proving effective in few-shot learning and domain shifts for clinical applications.