This article takes a look at the widely used machine learning application: computer vision. While it is used heavily in many market verticals, we will focus on its usage in the R&D-driven pharmaceutical and related industries.

To remove any ambiguity, machine vision is part of the larger computer vision domain. We use computer vision because we discuss algorithms to work with image data, as we will discuss here.

Zifo’s approach has been to acquire an in-depth knowledge of AI/ML applications/solutions already in use and focusing on the science. We are here to deliver solutions and bring innovations because that is the place from which transformational impacts derive.

So, let us take a critical look at the pharmaceutical and biotech industries’ state and related “science-centric” industries like Chemicals, Food, Fast Moving Consumer Goods, Agro-tech, and Oil & Gas’s use of this technology.  


What computer vision brings?

One pharmaceutical industry activity is assessing products and packaging to guarantee optimum quality and deliver traceability via the supply chain. Many other examples can be quoted, such as contamination measurement, cap integrity or code validation. This product tracking and quality assurance is not just a Pharma problem; it spans all the industries mentioned above – albeit perhaps in different terminologies.

Pharmaceutical companies oversee supply tracing and manufacturer data with all their products. Computer vision solutions can improve to identify that information and deliver it across networks during many steps of the shipping process.

The quality and quantity of the data have a precise effect on the consistency of the predictions provided by machine learning algorithms. Data is the fuel for data science, it has many features, such as instructions regarding medication or expiration dates, and this information is essential for distribution and consumption.

In brief, computer vision is one of the many machine learning applications that can bring efficiency benefits within strict quality demands. I can cite you our RPA (Robotic Process Automation) application named qcKen, a hyper automatization for Master Data Management, and an interface incorporated for values configuration. Our solution uses computer vision to help optimise the accuracy of the paper document ingestion.


What’s the future impact of computer vision?

Computer vision is growing fast and will impact all our futures as a part of improving all operations requiring human eyes — even activities in which human interface is needed, such as machine maintenance and manual quality assurance.

Currently, the primary use of computer vision is “digitising” forms and spreadsheets. Here, computer vision models can significantly accelerate text recognition and data capture whilst monitoring quality and “automating the human” interaction, e.g. in our solution qcKen.

We currently know how to robotically extract data from “paper” documents, making it accessible in a central system. Globally it helps digitise clinical trial documents, patient reports containing historical clinical data or historical medical records and even historical lab notebooks.

AI (Artificial Intelligence), and especially ML (Machine Learning), still needs human control since its mostly statistics and probabilities at the root. It would be reckless to let algorithms do sales target setting and business strategy by themselves without a human involved, and so there needs to be a a complementary flow and interaction.

For the future, the industry needs people to configure and experiment with these new technologies and optimise their use. In other words, computer vision is not meant to replace people but to augment and support their performance, as seen in image-based disease diagnostics case studies.

In conclusion, science-driven industries will modernise and deliver citizen health benefits by using AI/ML because everyone wants to ensure speed and accuracy is at the forefront of their research and development. With this in mind, a digital transformation is the only real option that can help them compete, innovate, and efficiently comply with legislation.

To find out more about how Zifo can help with AI, ML, Deep Learning, and scientific Data Sciences, please email me directly at

ZifoZifo RnD Solutions