OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Image & Video AI

Decimer Image Transformer

DECIMER Image Transformer is an open-source Python model that recognizes chemical structures in images and converts them into SMILES strings, using an EfficientNet V2 image encoder and a transformer for prediction.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Decimer Image Transformer — OpenViST project profile in Image & Video AI.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use DECIMER when the task is to turn a chemical structure drawing in an image into a SMILES string, including document digitization or code-based processing of an individual image.

Who it is for

  • Chemists digitizing structure drawings in scanned documents.
  • Researchers extracting chemical structures from scientific literature or patent images.
  • Python users who want to call a chemical image recognition function in code.

Documented uses

  • Convert an image of a chemical structure into a SMILES string.
  • Digitize chemical structure figures from literature, documents or patent images.
  • Use the documented Python package function to request a prediction from an image path.
Before you choose

Requirements & limits

The repository is public and lists an MIT license. Its README documents installation through Conda and pip.

  • The model architecture combines EfficientNet V2 image feature extraction with a transformer for SMILES prediction.
  • The documented example requires Python 3.10.0 in a Conda environment and uses the decimer package.
  • The README includes a separate hand-drawn model reference.
  • The repository points to decimer.ai as a separate way to experience DECIMER.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Install the Python package and convert one chemical image

Set up DECIMER in a Conda environment and use its documented Python function to request a SMILES string from an image.

Before you start

  • Conda
  • Python 3.10.0
  • A chemical structure image and its local file path

Steps

  1. Create a Conda environment named DECIMER with Python 3.10.0.

  2. Activate the DECIMER environment.

  3. Install the package with pip install decimer.

  4. Import predict_SMILES from DECIMER, pass it the image path, and print the returned value.

Expected outcome

The example returns a SMILES value for the supplied chemical image and prints it.

Guide limitations

  • The example documents one image path at a time and does not describe batch processing.
  • The README does not specify deployment requirements beyond the listed Python environment.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
15 October 2026
Original website title
Kohulan/DECIMER-Image_Transformer · Official homepage title · retained in its source language
How we review sourcesSuggest a correction