Out of distribution detection in deep learning-based scattering pattern classification

David Dannhauser; Paolo A. Netti; Filippo Causa

doi:10.1117/12.2670208

11 August 2023 Out of distribution detection in deep learning-based scattering pattern classification

David Dannhauser, Paolo A. Netti, Filippo Causa

Proceedings Volume 12622, Optical Methods for Inspection, Characterization, and Imaging of Biomaterials VI; 126220F (2023) https://doi.org/10.1117/12.2670208
Event: SPIE Optical Metrology, 2023, Munich, Germany

Abstract

Intrinsic biophysical cell properties hold an enormous potential for cell class and state classification in microfluidics, allowing to avoid the need of cost intensive fluorescence labelling. Several methods can accomplish cell identification, while convolutional neural networks show an outstanding performance compared to other state-of-the-art classification methods, regarding accuracy and speed. In fact, neural networks show high performance for known image class prediction but struggles when unknown (out of distribution) image classes need to be identified. In such a scenario no prior knowledge of the unknown cell class can be used for the model training, which inevitably results in image misclassification. In fact, to distinguish unknown cell classes, a neural network must first construct an in-distribution of known images to afterwards detect out of distribution as unknowns, which is also called open-set classification assumption. Ones, a new cell class is identified, the neural network can be retrained with the obtained knowledge to dynamically update its cell class database. This process can be simply repeated for each new detected cell class. We applied this open-set idea to scattering pattern snapshots of different classes of living cells obtained in microfluidics. Our outcome shows a proof-of-concept for open-set based convolutional neural network for cell image classification, which can be applied to a wide range of single cell classification approaches to reduce uncertainty in machine learning based technologies.

Conference Presentation

Citation Download Citation

David Dannhauser, Paolo A. Netti, and Filippo Causa "Out of distribution detection in deep learning-based scattering pattern classification", Proc. SPIE 12622, Optical Methods for Inspection, Characterization, and Imaging of Biomaterials VI, 126220F (11 August 2023); https://doi.org/10.1117/12.2670208

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available