Smart Engines Revolutionizes Code Quality Assessment with AI-Driven Tomography for Punched Card Input Validation

Specialists from the Russian AI company Smart Engines have made a breakthrough in the field of code review. Researchers have introduced a novel method for bug detection and quality assessment of code migration to physical media, utilizing artificial intelligence and computed tomography. This process is fully automated, eliminating the interface issues that often arise between the user and the computer.

Launching programs on punch cards is a costly endeavor, where the stakes are particularly high. Errors can occur during the preparation of the punch card set, such as perforations in the wrong locations, unpunched holes, or the mix-up of card order, which can invalidate the entire computation cycle and lead to a loss of valuable machine time. Operators strive to minimize such mistakes when transferring code to physical media, but even seasoned professionals are not immune to errors in programming. Thus, it’s essential to verify the quality of the code before executing a program on a specialized computing device.

The team at Smart Engines has proposed an innovative approach that incorporates various cutting-edge technologies, including X-ray tomography and computer vision, to tackle this pressing issue automatically. In the initial stage of the technological process, a batch of punch cards is bound and placed into a tomograph. After completing the tomographic measurement, the data is automatically corrected and reconstructed, resulting in a 3D model of the punch cards.

A key challenge is that to analyze the quality of the code, the punch cards must be identified, which necessitates aligning the individual cards in the stack by layers. The Smart Tomo Engine software solution includes a function that allows for the automatic alignment of flat objects, such as printed circuit boards or punch cards. Following the alignment, it is crucial to segment the holes on the punch cards and transfer this information to virtual punch cards, thus enabling automated comparison of the obtained models with a blueprint or text document.

«I urged my team to ensure that this technology was ready by April 1st. While the primary application of the developed method is quality control of code, it also holds significant potential for assessing the quality of multilayer printed circuit boards, synthetic materials in the form of crystalline films, and other objects with complex layered structures,” said Smart Engines’ CEO, Doctor of Technical Sciences Vladimir Arlazarov.