Artificial intelligence andCTCombining technology to tailor better imaging quality and efficiency

The software reconstruction algorithm is an important foundation for X-ray 3D tomography imaging technology. Good CT products should not only have excellent hardware conditions, but also be equipped with excellent reconstruction algorithms. After more than 20 years of development, Zeiss Xradia X-ray tomography imaging technology has been meticulously crafted in hardware and refined in software reconstruction algorithms, ensuring that the product system maintains mature and stable quality and has won the favor of a large number of users. In order to meet the pursuit of image quality and work efficiency by a large number of users, Zeiss has launched the Advanced Reconstruction Toolbox (ART) on the Xradia 3D X-ray Microscope (XRM) or Context Micro CT System, which can increase scanning speed by up to 10 times or significantly improve image quality at the same speed without sacrificing image quality, taking 3D X-ray tomography reconstruction technology to a new levelThe height.

Zeiss 3D X-ray Advanced Reconstruction (ART)includingOptiReconTheDeepRecon ProandPhaseEvolvemoduleEspecially the latest releaseDeepRecon ProandPhaseEvolveThe module adopts artificial intelligence （AI）Technology, compared to being based on'Filtered backprojection'Or standardFDKThe traditional reconstruction algorithm of the algorithm has achieved significant improvements in imaging speed and quality.

ZeissDeepRecon Pro

ZeissDeepRecon ProIt is an AI based reconstruction technology that can provide the most for various sample types 10 The advantage of doubling throughput or improving image quality saves a lot of scanning time. It is suitable for workflows involving semi repetitive and repetitive samples, as well as for individual samples. A user-friendly interface allows users to experience "one click" self training of machine learning network models, which can then be applied to sample like reconstruction.

Zeiss DeepRecon Pro is used for ceramic matrix composite (CMC) samples, achieving a 10 fold speed increase without sacrificing image quality. This provides higher temporal resolution for in-situ research. The left image shows standard reconstruction (FDK): scan time of 9 hours, 3001 projections; The middle image shows the standard reconstruction (FDK) with a scanning time of 53 minutes and 301 projections. The right image shows the Zeiss DeepRecon Pro with a scanning time of 53 minutes and 301 projections.

Zeiss DeepRecon Pro is used for 2.5D semiconductor intermediate layer packaging, achieving a 4x speed improvement without sacrificing image quality. The reconstruction results of DeepRecon Pro still show cracks of about 1um, and the signal-to-noise ratio is significantly improved. The left image shows standard reconstruction (FDK): scan time of 2 hours, 1201 projections; The middle image shows the standard reconstruction (FDK) with a scanning time of 30 minutes and 300 projections. The right image shows the Zeiss DeepRecon Pro with a scanning time of 30 minutes and 300 projections.

ZeissDeepRecon ProThe battery samples used in smartwatches have significantly improved image quality under the same scanning time, including both positive and negative electrode material image qualityThe left image isStandard refactoring; The image on the right shows ZeissDeepRecon ProThe scanning time is 6 hours.

ZeissPhaseEvolve

Zeiss PhaseEvolve is a post-processing algorithm for reconstructed data, which uses software algorithms to process boundary effects caused by phase contrast during low-density material photography, in order to improve the uniformity of imaging results and facilitate more accurate quantitative analysis of subsequent data segmentation, saving a lot of time in quantitative analysis.

Zeiss PhaseEnvolveApplied to drug powder samples. High resolution or low voltage imaging can cause the inherent image contrast of materials to be masked by phase effects. Zeiss PhaseEnvolveEffectively remove phase enhanced edges to enhance material contrast and improve image segmentation. The left image shows the standard refactoring; The image on the right shows PhaseEvolverestructure.

ARTModule scope of application:

The Zeiss Advanced Refactoring Toolbox improves the process of data acquisition and analysis, accelerates decision-making speed, and is suitable for research such as failure analysis of electronic semiconductors, earth sciences, pharmaceuticals, batteries, engineering materials, and 4D in-situ experiments. It is particularly suitable for multiple scanning tests of the same parameters in 4D in-situ research. The dilemma of image quality and sample scanning speed can be well solved through the Zeiss Advanced Refactoring Toolbox.

As a Zeiss Advanced Refactoring ToolboxARTOne of the users, Department of Earth Sciences, Utrecht University, NetherlandsMarkus OhlThe doctor said, "Cai SiDeepRecon ProProvided based onAThe simple yet powerful application of I and neural network technology allows users to easily achieve deep learning based applications without the need to understand deep learning techniquesXReconstruction of X-ray tomography. ”

Zeiss OptiReconDeepRecon ProandPhaseEvolvemoduleAll can be found in the existing ZeissXradia VersaSeries X-ray Microscope Context MicroCTUpgrade up. The Zeiss Customer Experience Center has been installed and upgraded. Interested new and old users are welcome to contact us to experience the new imaging effects brought by advanced AI technology reconstruction functions.