Computer Vision for Pharma

GxP-Compatible Quality Control Technology for Biologics, Cell and Gene Therapies, and Vaccines

Get started
training.py
inference.py
device.py
from mp_sdk import inference


inference_request = inference.get_request()


model_file = inference_request.model_file


my_model = load_model(model_file)


while inference_request.has_next_image():
    image = inference_request.next_image()
    # Run inference on image
    results = my_model.run(image)
    # Report results
    image.save_inference_results(results)

Introduction

Overview

AI software automating image analysis across multiple unit operations in the biopharmaceutical value chain.

mpVision is a hardware agnostic software layer added to any camera-equipped devices, ranging from microscopes to microflow imaging devices. All the software's accuracy, speed, security, and real-time processing requirements have been tested extensively by Top 20 Pharma on a global scale. The cloud-based technology connects with any imaging device currently implemented by the pharmaceutical company through either our API or web browser.

We enable researchers, data scientists, and developers to train their own AI to achieve real-time classification of any imaging data - from biologic morphology analysis to commercial contamination detection - for comprehensive quality control across R&D and manufacturing.

The software has been globally designed and tested by scientists at a Top 10 Pharma, and its features are mapped to the unmet needs of the pharmaceutical industry, including real-time product quality assurance, full agency compliance, faster go-to-market, and reduced operating costs.

Whether you are a biologist, a data scientist, or a quality manager, mpVision is designed for your quality needs in the lab and on the commercial production floor.

Tutorials

Create your AI models for your production line.

User Manual

Software functionality and usage guidelines.

SDK Documentation

Plug in custom AI models to mpVision ecosystem for training and inference.

API reference

Integrate your devices and build custom pipelines and monitoring functions.

Modicus Prime has performed side-by-side comparisons of mpVision with current softwares being used on the market, as was part of a triple bid process with a Top-10 Pharma. All the technology's accuracy (>99%), speed (0.6 seconds/-frame), and data processing (130K images in <24 hours) requirements have been tested extensively on a global scale. This unique multi-functional technology allows domain experts to train their own AI models on their confidential image data in-house, and these models can then be rolled out for real-time processing in a lab or a commercial production floor. It is a re-usable AI model generator that trains proprietary models across different biologics in a portfolio. These AI models are owned by pharma customers to bolster their competitive advantage.

Next
For GxP