---
title: "Smarter Image Labeling Boosts ResNet Accuracy by 23% Without Changing the Model"
subtitle: "A hierarchy-guided crowdsourcing framework that forces annotators to verify visual properties rather than pick a name, boosting agreement to 0.974."
url: "https://alphasignal.ai/news/smarter-image-labeling-boosts-resnet-accuracy-by-23-without-changing-the-model"
author: "AlphaSignal Newsroom"
author_url: "https://alphasignal.ai/editorial-team"
company: "AlphaSignal"
company_url: "https://alphasignal.ai/company/alphasignal"
published: "2026-04-15T00:00:00.000Z"
updated: "2026-08-23T20:43:04.821Z"
type: "Paper"
topics:
  - "Data"
  - "Image"
subtopics:
  - "Data Curation"
  - "Vision Language"
  - "Object Detection"
upvotes: 1700
source_url: "https://arxiv.org/abs/2604.14449"
comments: 0
truncated: true
---

# Smarter Image Labeling Boosts ResNet Accuracy by 23% Without Changing the Model

> A hierarchy-guided crowdsourcing framework that forces annotators to verify visual properties rather than pick a name, boosting agreement to 0.974.

Image datasets are the fuel for computer vision, yet the way we label them has barely changed since [ImageNet](https://arxiv.org/abs/2604.14449): show an annotator a picture, hand them a list of words, ask them to pick one. A new paper argues that this workflow quietly poisons the training signal, and proposes a crowdsourcing protocol that asks annotators about visual properties instead of names.

## The bear that is also a teddy, a cartoon, and a costume

The paper opens with two uncomfortable examples from popular benchmarks. In ImageNet, a single guitar image sits simultaneously under "Musical Instrument", "Guitar", and "Acoustic Guitar". In the Open Images Dataset, the "Brown Bear" category lumps together a real animal, a teddy bear (also tagged "Teddy Bear"), a stick-figure cartoon bear, and a person in a bear costume.

The authors trace both failures to the semantic gap problem: the mismatch between what can be extracted from pixels and how a human interprets those pixels in context. When annotators are handed only a word like "bear," the granularity of the label and the criteria for membership are left to personal taste. Three annotators produce three different label distributions, and a downstream classifier inherits the noise.

## Visual genus and visual differentia

The proposed methodology borrows a trick from classical taxonomy. Every category in a hierarchy is defined by two things: a **visual genus**, the properties it shares with its parent, and a **visual differentia**, the properties that separate it from its siblings. For "Goldfinch," the visual genus is "Finch" (used as the parent category), and the visual differentia is "Crimson face and yellow-and-black wings," which distinguishes it from sibling categories.

_Preview only. Full text requires an AlphaSignal Pro account._

## Sources

- [Crowdsourcing of Real-world Image Annotation via Visual Properties](https://arxiv.org/html/2604.14449)
- [Crowdsourcing of Real-world Image Annotation via Visual Properties (abstract)](https://arxiv.org/abs/2604.14449)
- [Prolific crowdsourcing platform](https://www.prolific.com/)