arXiv.itsnick.co

an independent field guide to 23 million papers from arXiv & OpenAlex — not affiliated with arXiv.org

Author

Dustin Morley

3 papers · 2010–2017

2010: 12011: 02012: 02013: 02014: 02015: 02016: 02017: 2
2010peak 2 in 20172017

Frequent co-authors

Papers

showing 1–3 of 3 RSS

Improving RANSAC-Based Segmentation through CNN Encapsulation

Dustin Morley, Hassan Foroosh

Signal Processing · Journal of International Crisis and Risk Communication Research · · cited 20× · DOI ↗ · more like this ·

In this work, we present a method for improving a random sample consensus (RANSAC) based image segmentation algorithm by encapsulating it within a convolutional neural network (CNN). The improvements are gained by gradient descent training on the set of pre-RANSAC filtering and thresholding operations using a novel RANSAC-based loss function, which is geared toward optimizing the strength of the correct model relative to the most convincing false model. Thus, it can be said that our loss function trains the network on metrics that directly dictate the success or failure of the final segmentation rather than metrics that are merely correlated to the success or failure. We demonstrate successful application of this method to a RANSAC method for identifying the pupil boundary in images from the CASIA-IrisV3 iris recognition data set, and we expect that this method could be successfully…

Simultaneous Detection and Quantification of Retinal Fluid with Deep Learning

Dustin Morley, Hassan Foroosh, Saad Shaikh +1 more

cs.CV · · arXiv ↗ · more like this ·

We propose a new deep learning approach for automatic detection and segmentation of fluid within retinal OCT images. The proposed framework utilizes both ResNet and Encoder-Decoder neural network architectures. When training the network, we apply a novel data augmentation method called myopic warping together with standard rotation-based augmentation to increase the training set size to 45 times the original amount. Finally, the network output is post-processed with an energy minimization algorithm (graph cut) along with a few other knowledge guided morphological operations to finalize the segmentation process. Based on OCT imaging data and its ground truth from the RETOUCH challenge, the proposed system achieves dice indices of 0.522, 0.682, and 0.612, and average absolute volume differences of 0.285, 0.115, and 0.156 mm$^3$ for intaretinal fluid, subretinal fluid, and pigment…

Cavity mode waves during terahertz radiation from rectangular Bi$_2$Sr$_2$CaCu$_2$O$_{8+δ}$ mesas

Richard A. Klemm, Erica R. LaBerge, Dustin R. Morley +3 more

cond-mat.supr-con · · arXiv ↗ · more like this ·

We re-examined the angular dependence of the radiation from the intrinsic Josephson junctions in rectangular mesas of Bi$_2$Sr$_2$CaCu$_2$O$_{8+δ}$, in order to determine if the cavity mode part of the radiation arises from waves across the width $w$ or along the length $\ell$ of the mesas, associated with ``hot spots'' [Wang {\it et al.}, Phys. Rev. Lett. {\bf 105}, 057002 (2010)]. We derived analytic forms for the angular dependence expected in both cases for a general cavity mode in which the width of the mesa corresponds to an integer multiple of one-half the wavelength of the radiation. Assuming the coherent radiation from the $ac$ Josephson current source and the cavity magnetic surface current density source combine incoherently, fits to the data of Kadowaki {\it et al.} [J. Phys. Soc. Jpn. {\bf 79}, 023703 (2010)] on a mesa with mean $\ell/w=5.17$ for both wave directions using…