Evaluating the Use of Edge Detection in Extracting Feature Size from Scanning Electrochemical Microscopy Images.
Lisa I StephensNicholas A PayneSebastian A SkaanvikDavid PolcariMatthias GeisslerJanine MauzerollPublished in: Analytical chemistry (2019)
The edge of a reactive or topographical feature is hard to estimate from feedback-based scanning electrochemical microscopy due to diffusional blurring, but is crucial to determining the accurate size and shape of these features. In this work, numerical simulations are used to demonstrate that the inflection point in a 1D line scan corresponds well to the true feature edge. This approach is then applied in 2D using the Canny algorithm to experimental images of two model substrates and a biological sample. This approach circumvents the need for aligning the imaged region between separate microscopy techniques, reveals hidden details embedded in SECM images, and allows individual features to be separated from their background more effectively.
Keyphrases
- deep learning
- label free
- high resolution
- convolutional neural network
- optical coherence tomography
- machine learning
- single molecule
- high speed
- gold nanoparticles
- electron microscopy
- high throughput
- computed tomography
- ionic liquid
- molecularly imprinted
- molecular dynamics
- magnetic resonance imaging
- magnetic resonance
- electron transfer
- real time pcr