This article first appeared in Digital Edge, The Edge Malaysia Weekly on May 10, 2021 - May 16, 2021

About five years ago, Samsung mobile phones made headlines when the batteries of specific models exploded. The unfortunate incidents were essentially caused by a tiny crack that was found on the batteries, which was difficult to spot with the naked eye.

Machine vision technology can prevent such occurrences, as it pairs camera lenses with artificial intelligence (AI) to carry out quality control inspections, says Edmund Yuen, CEO of Ideal Vision Integration Sdn Bhd.

“This is one of the main reasons an automated process is better than the human eye. Furthermore, an automated machine vision system would allow for traceability, where images and videos can be recorded and stored as backup and for future reference,” says Yuen.

Ideal Vision, which was founded in 2013, specialises in providing advanced machine vision solutions — both software and hardware — to businesses. Its largest clients are semiconductor companies, followed by medical, automotive and packaging companies. Recently, the company has been focusing its research and development (R&D) on automated optical inspection (AOI) systems, which is in demand by semiconductor companies.

These systems will eventually replace humans in a quality control factory line, as consistency and accuracy are what companies are looking for, says Yuen. They also cut down inspection time significantly, making them more attractive to multinational companies.

“Humans are still needed to check on the machines and learn how to control them. But at the rate factories and companies are making the shift to incorporate technology, human inspections will most likely be a thing of the past, replaced by automated machine vision systems,” he says.

One of the biggest challenges is getting companies to invest in the systems, which can cost anything from a few hundred thousand ringgit to a few million, says Yuen. He adds that typically, a company would calculate its return on investment (ROI) and how fast it can cover the cost of the purchase. With semiconductor companies, it may only take two to three years but with others, it would take a bit more time.

Recently, the company has been offering a leasing programme to loyal customers that have also been hit by the pandemic. With this programme, Ideal Vision will install the machine and software for the client to try out. After testing it for three to six months, they have the option of converting the lease to a purchase.

“This model has been widely seen with consumer goods, especially at the height of the pandemic. We may gradually change our business model to this but we need funding support to sustain this model. Once we are financially secure, we’ll look into this,” says Yuen.

The medical industry has also been using this technology to check medical instruments and devices, such as pacemakers and glucose checkers. Yuen says that recently, the company has seen demand from the glove and face mask sectors due to the pandemic.

“The public is concerned about the quality of these items and medical factories may be looking at machine vision systems to ensure the quality of their products meets the demand of customers.”

While Ideal Vision had been focusing its efforts on the local market, multinational companies that have a presence locally started exporting its solutions to their factories in the region over time. Ideal Vision now has a presence in Southeast Asia, China, Taiwan and, most recently, in the US and the UK.

Due to funding limitations, Yuen says Ideal Vision was not able to invest a lot in R&D to develop new products. The company then decided to shift its strategy to focus more on software developments, specifically exploring deep learning, which is an advanced version of machine learning. This also allows Ideal Vision to sell its software to companies that may already have the infrastructure for machine vision technologies.

The company is focusing on image processing and expression capabilities, says Yuen, leveraging its past experiences to develop a new machine vision software platform to be applied for inspections.

“Deep learning will be quite advanced for machine vision technologies because while we still need to set the parameters, we will just need to input data into the system to teach it right from wrong and eventually, it will learn by itself from the data collected,” Yuen explains.

Ideal Vision is also working with academic institutions such as Universiti Sains Malaysia, Universiti Malaysia Pahang and Universiti Tunku Abdul Rahman on technology innovations. Yuen says in recent years, collaborations with universities have been more focused on AI capabilities.

Tech talent in the machine vision space is scarce, says Yuen, as machine vision is not a major part of university courses and only part of a subject. This means that a great deal of the company’s time is spent on training fresh graduates, helping them understand machine learning and how to adapt to the company’s wants and needs.

“Even I didn’t know what machine vision was in university, until I did my internship. So now, we work closely with these universities to share our knowledge and experience, so students would know about this field,” he says.

“Gradually, we’re seeing some academic institutions putting a focus on machine vision. We also collaborate with students on their Master’s and PhD projects and from there, we can scout for talent.”

For the company to keep abreast of the latest technology developments, Yuen says the best way is to gain exposure overseas by attending conferences and workshops. Every year, Ideal Vision engineers are sent to attend technical training, workshops and conferences overseas but over the last two years, this has not been possible due to pandemic travel restrictions.

High-speed fibre connectivity and 5G are some of the aspects of MyDIGITAL that Ideal Vision is looking forward to utilising, says Yuen. This is because machine vision technology requires high speed data transfer for photos and videos collected through the machine. The company also employs cloud AI processing and these facilities will allow the company to provide an effective and cheaper solution to its customers.

“At present, our customers are still using local area network (LAN) cables because they are the fastest connectivity available for factories. I believe with the implementation of 5G, we will be able to do a lot of more complicated processing as it needs very high-speed network,” says Yuen.

“Our customers in China are already utilising 5G but since the facility is not yet available in our country, we can’t work on developing solutions for these customers. I believe we can create a very powerful machine vision application and so we are looking forward to seeing MyDIGITAL come to life.”

However, Yuen believes that the 10-year timeline of MyDIGITAL may be too long and should be reduced to five years. This is because technology is advancing at a fast pace nowadays and his concern is that the country may be left behind.

“Hopefully, our government will react to these upcoming trends and execute the plan to catch up with the pace of global technology. Companies in China and the US are already working on 6G technology and within two to three years, the concept and prototype might be ready.”

Ideal Vision won the award for the Most Innovative SME — Technology Solutions Provider at the SME & Entrepreneurs Business Award 2020 (SEBA 2020). Yuen attributes the win to its spirit of change and innovation.

“We have never stopped learning and growing our technological capabilities. In eight years, we have been able to transform a very small company to provide a flexible integration machine vision platform that benefits our customers,” he says.

Save by subscribing to us for your print and/or digital copy.

P/S: The Edge is also available on Apple's AppStore and Androids' Google Play.