Article
CommunityJev Models & Reimplementations
Every Pixel Is an Option: A System One Model for Medical Ultrasound Biometry
arXiv paper introducing Kev-Bio, a vision-language model that answers clinical measurement questions to place ultrasound calipers. It turns language-model decision representations into queries matched against dense ultrasound features, producing spatial probability maps for landmark coordinates.
- Format
- Article
- Reading time
- 2 min
- Published by
- Community
- Added
- Nov 2026
Highlights
- Kev-Bio is a vision-language model for ultrasound landmark detection guided by clinical measurement questions.
- Inspired by System One decision models, it generates queries for each caliper endpoint from language-model decision representations.
- Queries are matched against shared dense ultrasound features to produce spatial probability maps.
- On the nine-view FU/GU Biometry 2026 dataset it reports 22.50 pixels mean radial error and 15.38 clinical measurement MAE.
- It is trained end-to-end with landmark localization and clinical measurement losses; code and models are on GitHub.
Questions about this entry
Short answers drawn from the linked project or page.
Kev-Bio is a vision-language model described in an arXiv paper on medical ultrasound biometry. It uses clinical measurement questions to guide ultrasound landmark detection, then derives landmark coordinates and clinical measurements from spatial probability maps.
The paper says Kev-Bio is inspired by System One decision models: it uses decision representations of the language model to generate queries for each caliper endpoint, instead of predicting coordinates directly.
On the nine-view FU/GU Biometry 2026 dataset, Kev-Bio reports a mean radial error of 22.50 pixels and a clinical measurement mean absolute error of 15.38, outperforming the compared baselines in overall performance.
The paper states that code and models are available in the GitHub repository moeinheidari7829/Kev-Bio, which is linked from the arXiv abstract page for this work.
According to the paper, the model is trained end-to-end using both landmark localization losses and clinical measurement losses.
The provided material does not state a license, hardware, or runtime requirements for Kev-Bio. Those details would need to be checked in the linked GitHub repository and the paper itself.
Kev-Bio targets ultrasound landmark detection and clinical measurement across anatomical views. It is not a general chat or text-generation model, and the paper does not claim results outside ultrasound biometry tasks.
Reactions & coverage
Posts, threads, and videos about this entry from around the web.
X: Standard One SH (StandardThinking)
LinkedIn: What multimodal decision models actually change, and what they do not.
LinkedInJevDirectory.org
Hacker News: Decision Index 0.2.1 scores open reproductions of Jev on a frozen suite.
jackbravo4 pointsDecision Index 0.2.1 – 70 open reproductions of TypeSafe Jev's Decision ModelRead the thread on Hacker NewsHacker News: Pointer to a DiffusionGemma plus vLLM tutorial for a Jev-shaped classifier.
1 pointHow to build a Jev-style classifier with DiffusionGemma and vLLMRead the thread on Hacker NewsMore Jev resources like this
Open Jev-style decision model built on LiquidAI's LFM2.5-350M backbone, a clone of kev. Serves the System One API contract so the TypeSafe SDK works against a local server, returning probabilities for typed questions from one forward pass.
Jev Models & Reimplementations#open-models#python#local
Communityfranckverrot/lev
Open-weight vision-language judge that takes a task instruction plus 10 frames of a robot manipulation attempt and predicts success or failure with calibrated probabilities in a single forward pass. Released with inference code and a 20,628-prediction evaluation set.
Jev Models & ReimplementationsArticle#arxiv#open-models#evaluation
Community
A Hugging Face model repository for TechnoBaptist/MiniCPM-OCR-Jev, a MiniCPM-based OCR model carrying the Jev name. Its published chat template defines image and video vision tokens, XML-style tool calls, and think-tag reasoning output.
Jev Models & Reimplementations#open-models#github#research
Community
arXiv paper introducing Iris, a System One design that accepts image, text, or mixed candidates while leaving Jev's decision module untouched, with order-invariance, null-probability, and candidate-reuse results on a Multimodal-Mind2Web benchmark.
Jev Models & ReimplementationsArticle#arxiv#open-models#calibration
Community
arXiv paper testing whether the readout and adaptation choices used for direct selection still hold up in cross-request entailment reranking for JEV decision models. It combines a mean-span readout with selective backbone adaptation and reports accuracy on MC11, an average over eleven reranking configurations.
Jev Models & ReimplementationsArticle#arxiv#research#classification
Community
Apache-2.0 fine-tune of Qwen/Qwen3-ASR-0.6B published on Hugging Face that adds a Jev head for Dutch speech, tagged for data curation, speaker attributes, and audio quality. Ships custom modeling code and 782M BF16 parameters in one safetensors file.
Jev Models & Reimplementations#open-models#arxiv#research
Community
From the Jev guides
Original write-ups on what this entry is for.
From the Jev AI model community
Posts from builders shipping with Jev right now.