Quantifying defects in thin films using machine vision

Abstract

The sensitivity of thin-film materials and devices to defects motivates extensive research into the optimization of film morphology. This research could be accelerated by automated experiments that characterize the response of film morphology to synthesis conditions. Optical imaging can resolve morphological defects in thin films and is readily integrated into automated experiments but the large volumes of images produced by such systems require automated analysis. Existing approaches to automatically analyzing film morphologies in optical images require application-specific customization by software experts and are not robust to changes in image content or imaging conditions. Here, we present a versatile convolutional neural network (CNN) for thin-film image analysis which can identify and quantify the extent of a variety of defects and is applicable to multiple materials and imaging conditions. This CNN is readily adapted to new thin-film image analysis tasks and will facilitate the use of imaging in automated thin-film research systems.

A scalable thin-film defect quantify model under imbalanced regression and classification task based on computer vision

“…One reason is that labeling those images is labor-intensive, tedious, and repetitive work, whether ascertaining defect levels or calculating damage diameters. Fortunately, Project Ada is one of them, pursued by Nina et al [ 47 ] and collaborated with the Shanghai Institute of Ceramics, Chinese Academy of Sciences. Over twenty thousand thin-film defect images have been taken in this dataset, including multiple annealing methods and two different shooting light environments (Brightfield conditions for cracks, dewetting, parts, and scratches; Darkfield conditions for cracks, dewetting, no cracks, and no dewetting).…”

A self-driving laboratory designed to accelerate the discovery of adhesive materials

“…These images could be automatically processed with machine vision to help classify the bond failure mode or categorize morphological defects. 20 This information could be used to alert the researcher if an outlier occurred.…”

Defect chemistry of electrocatalysts for CO2 reduction

“…Owning to the crucial roles of defects sites in creating new functionalities of metal compounds, atomic-scale identification and quantification is very urgent. In addition to some common evaluation techniques such as Raman spectroscopy, electron-spin resonance spectroscopy, and infrared spectroscopy, etc., the recent rise of deep machine learning may also provide a feasible route to quantify defects ( Taherimakhsousi et al, 2020 ; Yang et al, 2021 ). According to the different microstructures, the defects of metal compounds can be divided into vacancies, grain boundaries, lattice defects and so on.…”