Tomra Food has introduced the Tomra 4C, an optical sorting machine designed to improve detection, accuracy, performance, efficiency and flexibility in nut and IQF applications. The system is the first from the company to run multiple forms of artificial intelligence in tandem, combining LUCAi deep learning with traditional machine learning. By leveraging AI models alongside updated sensing hardware, including the latest generation of pulsed LED sensors, the sorter is engineered to enhance identification of foreign materials as well as product and color defects. According to the company, the machine targets a false reject rate of less than 1 percent for nuts and less than 0.5 percent for IQF, while retaining the accuracy associated with Tomra chute sorting. The 4C is delivered pre-set and ready to integrate into new or existing lines, offering a small footprint and an interface intended to support everyday operation so lines can run at full speed with minimal adjustment.
Tomra positions the dual-AI approach as a way to increase adaptability when raw material conditions or sorting objectives change, aiming to stabilize quality and throughput across batches and shifts. The ability to deploy LUCAi and machine learning simultaneously is intended to help packers, processors and repackers quickly tune performance to evolving requirements in the nuts industry and beyond.
Dual AI architecture and detection performance
“The Tomra 4C will allow even more producers, processors and repackers to benefit from the significant developments AI, machine learning and deep learning bring to the food sorting industry. Combined, the technologies help increase detection of foreign materials, product and color defects, ensuring increased food safety and consistency across batches and shifts for optimized resource use and output,” said Harald Henriksen, executive vice president and head of Tomra Food.
Kristof Franckx, product manager for the Tomra 4C, highlighted the ejection accuracy enabled by the updated sensing and processing suite. “The Tomra 4C delivers serious performance improvements with the ability to achieve a lower than 1% false reject rate (<1% for nuts, and <0.5% for IQF), allowing our customers to set a new benchmark for the industry. This is possible thanks to the excellence in engineering that Tomra is known for combined with AI-powered processing and the latest generation of pulsed LED sensors. But most importantly, combining easy access and artificial intelligence delivers easy daily use for operators, saving time for operators and keeping the sorting line running at top speed.” According to the company, the 4C maintains the accuracy associated with its chute sorting machines while providing flexibility in AI deployment.
The system arrives pre-configured for rapid commissioning and is designed to fit into existing layouts. This approach is intended to reduce integration time and enable users to capitalize on AI-driven detection without extensive rework of current lines.
Company context and LUCAi expansion
The launch extends a series of Tomra Food product and application introductions since the start of 2025. “Tomra Food has been driving growth for Tomra Group for the past several quarters. We saw a strong order intake growth of 13% in 2024 and experienced a 15% revenue increase in the second quarter of 2025,” said Harald Henriksen. He added: “Our strong financial position combined with our strategic move to become closer to our customers in 2024 has created focus on priority product developments like the Tomra 4C, a machine that serves multiple industries with distinct challenges to solve.”
Tomra Food began 2025 with new LUCAi deep learning models for oranges, mandarins and lemons. First launched in 2018, LUCAi is now available for apples, blueberries, cherries and citrus, with more than 3,500 lanes sold globally. Additional LUCAi applications are under development for release in 2025 and 2026.
