Unlock instant, AI-driven research and patent intelligence for your innovation.

Transfer learning photoetching dead pixel detection method based on pre-trained deep convolutional neural network

A neural network and deep convolution technology, applied in the field of lithography dead pixel detection in integrated circuit design layout, which can solve problems such as insufficient precision and F1 score

Pending Publication Date: 2022-05-17
SHANGHAI INST OF OPTICS & FINE MECHANICS CHINESE ACAD OF SCI
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Existing transfer learning-based bad point detection method (see prior art 1, Kaibo Zhou, Kaifeng Zhang, Jie Liu, et al. An imbalance awarelithography hotspot detection method based on HDAM and pre-trained GoogLeNet[J].Measurement Science and Technology, 2021, 32: 125008.) The precision rate and F1 score are still insufficient

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Transfer learning photoetching dead pixel detection method based on pre-trained deep convolutional neural network
  • Transfer learning photoetching dead pixel detection method based on pre-trained deep convolutional neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0034] The present invention will be further described below in conjunction with embodiment and accompanying drawing, but should not limit protection scope of the present invention with this embodiment

[0035] The embodiment of the present invention adopts the dead point detection technology based on pre-trained GoogLeNet transfer learning (i.e. prior art 1) and SAMSUNG company's dead point detection technology based on deep convolutional neural network (i.e. prior art 2: MoojoonShin, Jee- Hyong Lee.Accurate lithography hotspot detection using deepconvolutional neural networks[J].J.Micro / Nanolith.MEMS MOEMS,2016,15(4):043507.), as a comparison object. In the embodiment of the present invention, the model is selected as the VGG13 network, and the graphic data set A used for model pre-training is the ImagNet data set. The integrated circuit layout data set B for model training is the training data set in the ICCAD2012 data set, and the layout graphics data set C is obtained aft...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a transfer learning photoetching dead pixel detection method based on a pre-trained deep convolutional neural network. According to the method, a pre-trained VGG convolutional neural network is used as a model, down-sampled layout graphic data is used as model input, and cross entropy loss is used as a loss function of model training, so that a model suitable for photoetching dead pixel detection is obtained through training. According to the method, the precision ratio and the F1 score of the defective pixel detection result can be effectively improved, and good comprehensive performance is obtained in recall ratio, precision ratio and F1 score.

Description

technical field [0001] The invention relates to a photolithography dead point detection technology, in particular to a photolithography dead point detection method based on a deep learning-based integrated circuit design layout. Background technique [0002] Photolithography is one of the key technologies for integrated circuit manufacturing. With the demand for high integration and better performance, the design size of semiconductors continues to shrink, and lithography manufacturability becomes one of the key issues. There are some areas in the integrated circuit layout where the lithography result is significantly different from the target pattern, which may cause a short circuit or an open circuit in the lithography result, thereby causing lithography hotspots. In the stage of IC layout design, hotspot detection and layout correction need to be carried out in turn, so as to avoid the occurrence of lithography dead spots. Dead point detection technology affects the cyc...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
IPC IPC(8): G06K9/62G06V10/774G06N3/04G06T7/00
CPCG06T7/0004G06T2207/30141G06N3/045G06F18/214
Inventor 廖陆峰李思坤王向朝
Owner SHANGHAI INST OF OPTICS & FINE MECHANICS CHINESE ACAD OF SCI