Structured illumination microscopy (SIM) reconstructs optically-sectioned images of a sample from multiple spatially-patterned wide-field images, but the traditional single non-patterned wide-field images are more inexpensively obtained since they do not require generation of specialized illumination patterns. In this work, we translated wide-field fluorescence microscopy images to optically-sectioned SIM images by a Pix2Pix conditional generative adversarial network (cGAN). Our model shows the capability of both 2D cross-modality image translation from wide-field images to optical sections, and further demonstrates potential to recover 3D optically-sectioned volumes from wide-field image stacks. The utility of the model was tested on a variety of samples including fluorescent beads and fresh human tissue samples.

No abstract

Digital pathology, or the practice of acquiring, managing, and interpreting high-resolution digital images from glass pathology slides, holds much promise in precision medicine, potentially transforming diagnosis and prognosis based on computational image biomarkers derived from digital tissue images. However, for all its promise, digital imaging in pathology has not yet become an integral part of the clinical workflow as it has in radiology due to high cost, workflow disruptions, burdensome data sizes and IT requirements, and additional dedicated personnel requirements. Consequently, pathology retains the 150-year-old analog workflow, and the vast majority of slides used in clinical diagnosis are never digitized. Furthermore, there is a missed opportunity to capture the image information and associated data on search processes that led to the clinical diagnosis, which could serve as the foundation for computational clinical decision support. This paper describes an approach for slide digitization during clinical review using a camera attached to a standard brightfield pathology microscope. While a pathologist reviews a glass slide using the eyepiece oculars, the continuously running camera digitizes a complete record of the slide review, resulting in multi-resolution slide images and spatiotemporal saliency maps of the slide review. Unlike other approaches, the pathologist does not stop to review the video stream or monitor the acquisition of video frames but performs the diagnostic review at the microscope using the standard clinical protocol. This hybrid analog-digital approach combines the benefits of digital slide analysis, including annotation, computation, and the ability to confirm the completeness and quality of the glass slide review with the ease of using the microscope for primary diagnosis. Furthermore, a record of the pathologist's attention during the review, including their search path, magnification level, and dwell times at each location on the slide, is obtained. In the future, this approach could enable the development and application of new and emerging computational decision-support algorithms in real-time to provide feedback to the pathologist, reduce diagnostic errors, and improve disease diagnosis and prognosis.

-Given the fact that most hearing-impaired listeners have low-frequency residual hearing, the present work assessed the effect of applying commonly-used singlechannel noise-reduction (NR) algorithms to improve the intelligibility of low-pass filtered speech, which simulates the effect of understanding speech with low-frequency residual hearing of hearing-impaired patients. In addition, this study was performed with Mandarin speech, which is characterized by its significant contribution of information present in (lowfrequency dominated) vowels to speech intelligibility. Mandarin sentences were corrupted by steady-state speechshaped noise and processed by four types (i.e., subspace, statistical-modeling, spectral-subtractive, and Wiener-filtering) of single-channel NR algorithms. The processed sentences were played to normal-hearing listeners for recognition. Experimental results showed that existing single-channel NR algorithms were unable to improve the intelligibility of lowpass filtered Mandarin sentences.1 Wiener-filtering had the least negative influence to the intelligibility of low-pass filtered speech among the four types of single-channel NR algorithms examined.

No abstract