Nuclei mymargin segmentation is a fundamental task for various computational pathology applications including nuclei morphology analysis, cell type classification, and cancer grading. Deep learning has emerged as a powerful approach to segmenting nuclei but the accuracy of convolutional neural networks (CNNs) depends on the volume and the quality of labeled histopathology data for training. In particular, conventional CNN-based approaches lack structured prediction capabilities, which are required to distinguish overlapping and clumped nuclei. Here, we present an approach to nuclei segmentation that overcomes these challenges by utilizing a conditional generative adversarial network (cGAN) trained with synthetic and real data. We generate a large dataset of H&E training images with perfect nuclei segmentation labels using an unpaired GAN framework. This synthetic data along with real histopathology data from six different organs are used to train a conditional GAN with spectral normalization and gradient penalty for nuclei segmentation. This adversarial regression framework enforces higher-order spacial-consistency when compared to conventional CNN models. We demonstrate that this nuclei segmentation approach generalizes across different organs, sites, patients and disease states, and outperforms conventional approaches, especially in isolating individual and overlapping nuclei.

To realize the full potential of deep learning for medical imaging, large annotated datasets are required for training. Such datasets are difficult to acquire due to privacy issues, lack of experts available for annotation, underrepresentation of rare conditions and poor standardization. The lack of annotated data has been addressed in conventional vision applications using synthetic images refined via unsupervised adversarial training to look like real images. However, this approach is difficult to extend to general medical imaging because of the complex and diverse set of features found in real human tissues. We propose a novel framework that uses a reverse flow, where adversarial training is used to make real medical images more like synthetic images, and clinically-relevant features are preserved via selfregularization. These domain-adapted synthetic-like images can then be accurately interpreted by networks trained on large datasets of synthetic medical images. We implement this approach on the notoriously difficult task of depth-estimation from monocular endoscopy which has a variety of applications in colonoscopy, robotic surgery and invasive endoscopic procedures. We train a depth estimator on a large dataset of synthetic images generated using an accurate forward model of an endoscope and an anatomically-realistic colon. Our analysis demonstrates that the structural similarity of endoscopy depth estimation in a real pig colon predicted from a network trained solely on synthetic data improved by 78.7% by using reverse domain adaptation.

We present an axisymmetric computational model to study the heating processes of gold nanoparticles, specifically nanorods, in aqueous medium by femtosecond laser pulses. We use a two-temperature model for the particle, a heat diffusion equation for the surrounding water to describe the heat transfer processes occurring in the system, and a thermal interface conductance to describe the coupling efficiency at the particle/water interface. We investigate the characteristic time scales of various fundamental processes, including lattice heating and thermal equilibration at the particle/surroundings interface, the effects of multiple laser pulses, and the influence of nanorod orientation relative to the beam polarization on energy absorption. Our results indicate that the thermal equilibration at the particle/water interface takes approximately 500 ps, while the electron-lattice coupling is achieved at approximately 50 ps when a 48×14 nm gold nanorod is heated to a maximum temperature of 1270 K with the application of a laser pulse having 4.70 J/m2 average fluence. Irradiation by multiple pulses arriving at 12.5 ns time intervals (80 MHz repetition rate) causes a temperature increase of no more than 3 degrees during the first few pulses with no substantial changes during the subsequent pulses. We also analyze the degree of the nanorods’ heating as a function of their orientation with respect to the polarization of the incident light. Lastly, it is shown that the temperature change of a nanorod can be modeled using its volume equivalent sphere for femtosecond laser heating within 5–15% accuracy.