Visual In-Context Learning for Few-Shot Eczema Segmentation

Neelesh Kumar; Oya Aran; Venugopal Vasudevan

RECORD

Visual In-Context Learning for Few-Shot Eczema Segmentation

Neelesh Kumar, Oya Aran, Venugopal Vasudevan

Computer Vision and Pattern Recognition · Submitted 28 Sep 2023

abstract compiled · 2026-09-23 · no language-model review

The authors' main claim Our result highlights the importance of visual in-context learning in developing faster and better solutions to skin imaging tasks.

how it is stated · asserted 0.94 read the record →

Comment (0) Read on arXiv ↗

Cite ▾

Claim paper

The compiled record

The paper's own sentences, arranged. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. Numbers in the tables are the authors', checked by code. No sentence in this record was written by a model.

Open the compiled record

Compiled by classifier-1.13.0 via TypeSafe API, question bank abstract-p1-20260923, 2026-09-23: 1 request, 7,581 input tokens.

What the abstract itself states, as probabilities: claims first 0.06 · states limitation 0.86 · human participants 0.22 · controlled comparison 0.90 · randomized assignment 0.01 · states quantitative result 0.99 · compares with prior results 0.98 · mentions code or data availability 0.03.

The abstract, sentence by sentence

arXiv listing abstract v1. Results and the main claim are underlined; show every role.

Automated diagnosis of eczema from digital camera images is crucial for developing applications that allow patients to self-monitor their recovery. An important component of this is the segmentation of eczema region from such images. Current methods for eczema segmentation rely on deep neural networks such as convolutional (CNN)-based U-Net or transformer-based Swin U-Net. While effective, these methods require high volume of annotated data, which can be difficult to obtain. Here, we investigate the capabilities of visual in-context learning that can perform few-shot eczema segmentation with just a handful of examples and without any need for retraining models. Specifically, we propose a strategy for applying in-context learning for eczema segmentation with a generalist vision model called SegGPT. When benchmarked on a dataset of annotated eczema images, we show that SegGPT with just 2 representative example images from the training dataset performs better (mIoU: 36.69) than a CNN U-Net trained on 428 images (mIoU: 32.60). We also discover that using more number of examples for SegGPT may in fact be harmful to its performance. Our result highlights the importance of visual in-context learning in developing faster and better solutions to skin imaging tasks. Our result also paves the way for developing inclusive solutions that can cater to minorities in the demographics who are typically heavily under-represented in the training data.

background contribution method result interpretation main claim

What the classifier reports for each sentence. The number is the probability it gave the label.
#	sentence	role	the authors' own work	states the main conclusion	strength
s0	Automated diagnosis of eczema from digital camera images is crucial for developing applications that allow patients to self-monitor their recovery.	background 1.00	0.05	0.03	asserted 1.00
s1	An important component of this is the segmentation of eczema region from such images.	background 1.00	0.05	0.02	asserted 1.00
s2	Current methods for eczema segmentation rely on deep neural networks such as convolutional (CNN)-based U-Net or transformer-based Swin U-Net.	background 1.00	0.04	0.02	asserted 1.00
s3	While effective, these methods require high volume of annotated data, which can be difficult to obtain.	gap 0.59	0.06	0.02	asserted 0.87
s4	Here, we investigate the capabilities of visual in-context learning that can perform few-shot eczema segmentation with just a handful of examples and without any need for retraining models.	contribution 0.99	0.54	0.10	hedged possibility 0.39
s5	Specifically, we propose a strategy for applying in-context learning for eczema segmentation with a generalist vision model called SegGPT.	method 0.50	0.83	0.06	asserted 0.99
s6	When benchmarked on a dataset of annotated eczema images, we show that SegGPT with just 2 representative example images from the training dataset performs better (mIoU: 36.69) than a CNN U-Net trained on 428 images (mIoU: 32.60).	result 1.00	0.97	0.50	asserted 1.00
s7	We also discover that using more number of examples for SegGPT may in fact be harmful to its performance.	result 1.00	0.96	0.14	hedged possibility 0.94
s8	Our result highlights the importance of visual in-context learning in developing faster and better solutions to skin imaging tasks.	interpretation 1.00	0.64	0.64	asserted 0.94
s9	Our result also paves the way for developing inclusive solutions that can cater to minorities in the demographics who are typically heavily under-represented in the training data.	interpretation 0.88	0.32	0.12	hedged possibility 0.75

What kind of paper the classifier reports this is: method or tool 0.91 · evidence from observation or experiment 0.96.

What was examined, and what was not

sources read

arXiv listing abstract v1

not examined

full text
figures and tables
references and citing sentences
relations to other papers

levels run

P abstract

not run

S sections
T body sentences
C citations
R cross-paper relations

Every quoted sentence on this record is the authors'. Every label is from Pith's closed vocabulary, with the probability shown beside it. Every number in the measurements table is the authors', checked by code. Record as JSON

The paper

Original abstract

Automated diagnosis of eczema from digital camera images is crucial for developing applications that allow patients to self-monitor their recovery. An important component of this is the segmentation of eczema region from such images. Current methods for eczema segmentation rely on deep neural networks such as convolutional (CNN)-based U-Net or transformer-based Swin U-Net. While effective, these methods require high volume of annotated data, which can be difficult to obtain. Here, we investigate the capabilities of visual in-context learning that can perform few-shot eczema segmentation with just a handful of examples and without any need for retraining models. Specifically, we propose a strategy for applying in-context learning for eczema segmentation with a generalist vision model called SegGPT. When benchmarked on a dataset of annotated eczema images, we show that SegGPT with just 2 representative example images from the training dataset performs better (mIoU: 36.69) than a CNN U-Net trained on 428 images (mIoU: 32.60). We also discover that using more number of examples for SegGPT may in fact be harmful to its performance. Our result highlights the importance of visual in-context learning in developing faster and better solutions to skin imaging tasks. Our result also paves the way for developing inclusive solutions that can cater to minorities in the demographics who are typically heavily under-represented in the training data.

Human review

How Pith checks papers → · Full audit of this paper →

How the literature uses it

What this paper does with the arXiv papers it cites

Each citing sentence is taken from the citing paper's LaTeX source and typed by Pith's classifier. A label appears at probability 0.8 or more (0.95 for support), a contest at 0.7 or more; the quoted words are the citing authors'. How each label was checked · JSON

What it cites · 6 sentences without a confident label

Uses the method or tool of 2 papers

The core of our method is a visual in-context learning model called SegGPT [ref] .

SegGPT: Segmenting Everything In Context · Methods · 0.99
The model in SegGPT employs a vanilla vision transformer (ViT) [ref] as the encoder consisting of stacked transformer blocks.

An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale · SegGPT Training · 0.93

Cites as support 3 papers

Our result adds to the mounting evidence that learning task-agnostic features on large diverse datasets with high-capacity models eliminates the need for performing any training or finetuning for downstream tasks [ref] .

Summary of ChatGPT-Related Research and Perspective Towards the Future of Large Language Models · Discussion · 1.00
Our result adds to the mounting evidence that learning task-agnostic features on large diverse datasets with high-capacity models eliminates the need for performing any training or finetuning for downstream tasks [ref] .

Segment Anything · Discussion · 1.00
Our result adds to the mounting evidence that learning task-agnostic features on large diverse datasets with high-capacity models eliminates the need for performing any training or finetuning for downstream tasks [ref] .

DINOv2: Learning Robust Visual Features without Supervision · Discussion · 1.00

Cites as background 6 papers

The idea is similar to that in NLP: a high-capacity network such as a Vision Transformer (ViT) [ref] is trained on a large corpus of images using Masked Image Modeling (MIM) [ref] .

An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale · Introduction · 0.89
A growing body of work is now attempting to replicate this success of NLP for vision tasks using a paradigm known as visual in-context learning [ref] .

What Makes Good Examples for Visual In-Context Learning? · Introduction · 0.85
A growing body of work is now attempting to replicate this success of NLP for vision tasks using a paradigm known as visual in-context learning [ref] .

Segment Anything · Introduction · 0.85

Discussion

Know something about this paper? Leave a comment, or highlight any passage above to comment on it.