Hardware Trojan horse detection method based on deep learning

Abstract

The invention discloses a hardware Trojan horse detection method based on deep learning. The method mainly solves the problems that an existing method is high in cost, high in omission ratio and low in efficiency. According to the scheme, the method comprises the steps of collecting image sets, and constructing two training sets; collecting a to-be-detected image set; training the residual channelattention network by using the first training set, training the cyclic consistency generative adversarial network by using the second training set, and sequentially sending the microscopic images inthe to-be-detected image set to the trained residual channel attention network and the cyclic consistency generative adversarial network to obtain images which are homologous with the mother set microscopic images; enhancing the images homologous with the mother set microcosmic images and the corresponding mother set microcosmic images, and carrying out binarization segmentation and denoising on the enhanced images; connected region marking being carried out on the denoised image, exclusive-OR operation being carried out on the denoised image, and the region with the operation result being 1 being a hardware Trojan horse. The method has the advantages of higher detection precision, higher speed and simpler operation, and can be used for preparing an integrated circuit chip.

Description

technical field [0001] The invention belongs to the technical field of image processing, and further relates to a hardware Trojan horse detection method, which can be used in the preparation of integrated circuit chips. Background technique [0002] With the rapid development of integrated circuit chip design and manufacturing technology, the increasing popularity of third-party technical services reduces the cost of chip manufacturers on the one hand, shortens the cycle of their entry into the market, and on the other hand increases the risk of chip security performance. Reduced reliability of integrated circuits. Under the conditions of economic globalization, it is easy for malicious parties to add some extra malicious circuit logic units that do not belong to the original design specifications during the process from design to manufacturing of integrated circuit chips, which is also called "hardware Trojan horse". [0003] In order to reduce the risk of using chips to c...

AI Technical Summary

Problems solved by technology

However, in the actual detection process, there are two problems in the hardware Trojan horse detection method based on electrical signals: one is that it is difficult to be actually detected due to the high concealment and low activation of the hardware Trojan horse; the other is that it is difficult to realize the full coverage detection of the chip

On the other hand, microscopic images of chips without hardware Trojans must exist in the sample chips using this method, otherwise, it is impossible to effectively classify and finally identify whether the chips carry hardware Trojans

Images