TY - JOUR
T1 - Deep learning optimization for small object classification in lensfree holographic microscopy
AU - Potter, Colin J.
AU - Sreevatsan, Shriniketh
AU - McLeod, Euan
N1 - Publisher Copyright:
© 2024 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement.
PY - 2024/9/23
Y1 - 2024/9/23
N2 - Lensfree holographic microscopy is a compact and cost-effective modality for imaging large fields of view with high resolution. When combined with automated image processing, it can be used for biomolecular sensing where biochemically functionalized micro- and nano-beads are used to label biomolecules of interest. Neural networks for image feature classification provide faster and more robust sensing results than traditional image processing approaches. While neural networks have been widely applied to other types of image classification problems, and even image reconstruction in lensfree holographic microscopy, it is unclear what type of network architecture performs best for the types of small object image classification problems involved in holographic-based sensors. Here, we apply a shallow convolutional neural network to this task, and thoroughly investigate how different layers and hyperparameters affect network performance. Layers include dropout, convolutional, normalization, pooling, and activation. Hyperparameters include dropout fraction, filter number and size, stride, and padding. We ultimately achieve a network accuracy of ∼83%, and find that the choice of activation layer is most important for maximizing accuracy. We hope that these results can be helpful for researchers developing neural networks for similar classification tasks.
AB - Lensfree holographic microscopy is a compact and cost-effective modality for imaging large fields of view with high resolution. When combined with automated image processing, it can be used for biomolecular sensing where biochemically functionalized micro- and nano-beads are used to label biomolecules of interest. Neural networks for image feature classification provide faster and more robust sensing results than traditional image processing approaches. While neural networks have been widely applied to other types of image classification problems, and even image reconstruction in lensfree holographic microscopy, it is unclear what type of network architecture performs best for the types of small object image classification problems involved in holographic-based sensors. Here, we apply a shallow convolutional neural network to this task, and thoroughly investigate how different layers and hyperparameters affect network performance. Layers include dropout, convolutional, normalization, pooling, and activation. Hyperparameters include dropout fraction, filter number and size, stride, and padding. We ultimately achieve a network accuracy of ∼83%, and find that the choice of activation layer is most important for maximizing accuracy. We hope that these results can be helpful for researchers developing neural networks for similar classification tasks.
UR - http://www.scopus.com/inward/record.url?scp=85205003810&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85205003810&partnerID=8YFLogxK
U2 - 10.1364/OE.527353
DO - 10.1364/OE.527353
M3 - Article
AN - SCOPUS:85205003810
SN - 1094-4087
VL - 32
SP - 35062
EP - 35081
JO - Optics Express
JF - Optics Express
IS - 20
ER -