---
title: "Logo Vectorization Benchmark: 15 Controlled Tests | Vectorize"
description: "See how Vectorize performed on five logos at 512px, 128px and with synthetic shading, including method, region IoU, topology, artifacts and failures."
canonical: "https://vectorize-logo.com/benchmarks/logo-vectorization/"
last_updated: "2026-10-14"
markdown_url: "https://vectorize-logo.com/benchmarks/logo-vectorization.md"
---

# Five logos. Fifteen controlled inputs. Failures included.

HTML version: https://vectorize-logo.com/benchmarks/logo-vectorization/ · Markdown: https://vectorize-logo.com/benchmarks/logo-vectorization.md · Site index for agents: https://vectorize-logo.com/llms.txt

> See how Vectorize performed on five logos at 512px, 128px and with synthetic shading, including method, region IoU, topology, artifacts and failures.

This test asks a narrow question: how closely does the Vectorize prototype recover the color regions and topology of clean logo rasterizations at two sizes, plus a lightly shaded variant?

- **Run**: September 10, 2026
- **Runner snapshot**: `344ddaf`
- **Cases**: 5 logos · 15 inputs
- **Dataset role**: Development, not independent

These cases were used to improve the prototype. They do not prove general quality or superiority over another product. The inputs are controlled PNGs derived from online reference SVGs—not camera images or downloaded JPEGs.

## The tracer sees only pixels.

Reference vectors are used to create and evaluate each test. They are never passed to the reconstruction pipeline.

01 **Raster input**

![Controlled Slack PNG input](https://vectorize-logo.com/benchmark/examples/slack-input-512.webp)

Pixels only. The reference SVG is not passed to the tracer.

02 **Detect regions**

![Automatically detected color regions](https://vectorize-logo.com/benchmark/examples/process-regions.png)

Connected areas become a compact, automatic color palette.

03 **Trace paths**

![SVG paths reconstructed from the detected regions](https://vectorize-logo.com/benchmark/examples/slack-output-512.svg)SVG

Each region is rebuilt as scalable geometry and flat fills.

04 **Render & measure**

**99.34%**: region IoU Topology matched

The SVG is rendered back to pixels for the same comparison.

One clean 512px Slack run shown. The same automatic pipeline is applied to every benchmark input without manual path editing.

- [Raw metrics (JSON)](https://vectorize-logo.com/benchmark/metrics.json)
- [Source URLs and hashes (JSON)](https://vectorize-logo.com/benchmark/source-manifest.json)
- [Independent evaluation protocol](https://vectorize-logo.com/benchmark/evaluation-protocol.md)

## Resolution is part of the result.

The left panel is the controlled raster input. The two output panels embed the generated SVG files directly—not screenshots of them—and let you download the same vectors. The 128px trace is enlarged so lost geometry remains easy to inspect.

### Adidas

Lettering, sharp corners and counters

[Reference source](https://commons.wikimedia.org/wiki/File:Adidas_Logo.svg)

![Adidas controlled 512-pixel PNG input](https://vectorize-logo.com/benchmark/examples/adidas-input-512.webp)

Input · 512px PNG

![Adidas rendered vector output traced from the 512-pixel input](https://vectorize-logo.com/benchmark/examples/adidas-output-512.svg)

Output · 512px source [SVG](https://vectorize-logo.com/benchmark/examples/adidas-output-512.svg)

![Adidas rendered vector output traced from the 128-pixel input](https://vectorize-logo.com/benchmark/examples/adidas-output-128.svg)

Output · 128px source [SVG](https://vectorize-logo.com/benchmark/examples/adidas-output-128.svg)

Good at 512px. At 128px the mark remains recognizable, but corners soften and the lettering loses precision.

### Slack

Four colors, separated parts and curves

[Reference source](https://commons.wikimedia.org/wiki/File:Slack_icon_2019.svg)

![Slack controlled 512-pixel PNG input](https://vectorize-logo.com/benchmark/examples/slack-input-512.webp)

Input · 512px PNG

![Slack rendered vector output traced from the 512-pixel input](https://vectorize-logo.com/benchmark/examples/slack-output-512.svg)

Output · 512px source [SVG](https://vectorize-logo.com/benchmark/examples/slack-output-512.svg)

![Slack rendered vector output traced from the 128-pixel input](https://vectorize-logo.com/benchmark/examples/slack-output-128.svg)

Output · 128px source [SVG](https://vectorize-logo.com/benchmark/examples/slack-output-128.svg)

Good in both clean sizes. The four regions stay separated, with less regular curves in the smaller trace.

### Wikimedia

Circles, symmetry and three colors

[Reference source](https://commons.wikimedia.org/wiki/File:Wikimedia-logo.svg)

![Wikimedia controlled 512-pixel PNG input](https://vectorize-logo.com/benchmark/examples/wikimedia-input-512.webp)

Input · 512px PNG

![Wikimedia rendered vector output traced from the 512-pixel input](https://vectorize-logo.com/benchmark/examples/wikimedia-output-512.svg)

Output · 512px source [SVG](https://vectorize-logo.com/benchmark/examples/wikimedia-output-512.svg)

![Wikimedia rendered vector output traced from the 128-pixel input](https://vectorize-logo.com/benchmark/examples/wikimedia-output-128.svg)

Output · 128px source [SVG](https://vectorize-logo.com/benchmark/examples/wikimedia-output-128.svg)

Good in both sizes. At 128px, circles and hard corners show a small visible approximation.

### Starbucks

Thin details and enclosed white regions

[Reference source](https://en.wikipedia.org/wiki/File:Starbucks_Corporation_Logo_2011.svg)

![Starbucks controlled 512-pixel PNG input](https://vectorize-logo.com/benchmark/examples/starbucks-input-512.webp)

Input · 512px PNG

![Starbucks rendered vector output traced from the 512-pixel input](https://vectorize-logo.com/benchmark/examples/starbucks-output-512.svg)

Output · 512px source [SVG](https://vectorize-logo.com/benchmark/examples/starbucks-output-512.svg)

![Starbucks rendered vector output traced from the 128-pixel input](https://vectorize-logo.com/benchmark/examples/starbucks-output-128.svg)

Output · 128px source [SVG](https://vectorize-logo.com/benchmark/examples/starbucks-output-128.svg)

The 512px result is similar but needs work in internal details. The 128px trace is not suitable for professional use.

### Google 2015

Multicolor lettering and open counters

[Reference source](https://commons.wikimedia.org/wiki/File:Google_2015_logo.svg)

![Google 2015 controlled 512-pixel PNG input](https://vectorize-logo.com/benchmark/examples/google-input-512.webp)

Input · 512px PNG

![Google 2015 rendered vector output traced from the 512-pixel input](https://vectorize-logo.com/benchmark/examples/google-output-512.svg)

Output · 512px source [SVG](https://vectorize-logo.com/benchmark/examples/google-output-512.svg)

![Google 2015 rendered vector output traced from the 128-pixel input](https://vectorize-logo.com/benchmark/examples/google-output-128.svg)

Output · 128px source [SVG](https://vectorize-logo.com/benchmark/examples/google-output-128.svg)

Good at 512px. The 128px trace stays recognizable, while stroke weight and letter geometry become less precise.

## Region overlap is a diagnostic, not a grade.

Mean color-region IoU is the intersection divided by the union of each rendered color region, averaged across colors. It is not an overall quality percentage and does not replace visual review.

Clean-input measurements by logo and source size

| Logo | Colors | Region IoU · 512 | Region IoU · 128 | Topology | SVG · 512 |
| --- | --- | --- | --- | --- | --- |
| Adidas | 1   | 98.74% | 94.52% | Matched | 4,940 B |
| Slack | 4   | 99.34% | 97.38% | Matched | 3,763 B |
| Wikimedia | 3   | 99.49% | 98.12% | Matched | 2,269 B |
| Starbucks | 2   | 97.75% | 89.44% | Did not match | 26,179 B |
| Google 2015 | 4   | 96.71% | 88.69% | Matched | 4,424 B |

### Geometry and rendering

Across clean 512px inputs, mean region IoU ranges from 96.71% to 99.49%. At 128px the range widens to 88.69%–98.12%. A high silhouette score can still conceal incorrect internal detail.

### Topology

Components and holes match the reference for four of five logo families at both clean sizes. Starbucks does not match in either size; features smaller than four pixels in the 4× evaluation are ignored.

### Color

Clean-input palettes are recovered exactly in these cases. On shaded inputs, the correct number of colors remains, but several representative colors become darker and white can become gray.

### Complexity

The clean 512px SVGs range from 2,269 to 26,179 bytes and from 4 to 36 traced regions. This is descriptive complexity, not a smaller-is-always-better target.

## The failed cases shaped the product.

The first experiment forced two colors, which failed on the one-color Adidas case and distorted multicolor palettes. The revised prototype estimates its palette, seeds disconnected regions, treats flat artwork separately and exports disconnected color shapes without a global base layer.

Starbucks at 128px improved from 71.03% to 89.44% mean region IoU, yet the face, star, hair and trademark detail remain visibly unreliable. The correct product response is to warn and review, not to call that result professional.

## Promising for simple logos, incomplete for small detail.

The independent evaluation protocol is now frozen separately from this development set. The next run must use unseen source families and add real JPEG compression, opaque backgrounds and more low-resolution lettering. Until those results exist, Vectorize should keep its narrow promise and require human review.

- [Try the converter](https://vectorize-logo.com/)
- [About the maker and method](https://vectorize-logo.com/about/)

The Adidas, Slack, Wikimedia, Starbucks and Google marks remain the property of their respective owners. They appear here only as attributed benchmark samples, not as Vectorize branding or endorsement.