A simple coordinate transformation method for quickly locating the features of interest in TEM samples.

We developed a simple coordinate transformation method for quickly locating features of interest (FOIs) of samples in transmission electron microscope (TEM). The method is well suited for conducting sample searches in aberration-corrected scanning/transmission electron microscopes (S/TEM), where the survey can be very time-consuming because of the limited field of view imposed by the highly excited objective lens after fine-tuning the aberration correctors. For implementation, a digital image of the sample and the TEM holder was captured using a simple stereo-optical microscope. Naturally presented geometric patterns on the holder were referenced to construct a projective transformation between the electron and optical coordinate systems. The test results demonstrated that the method was accurate and required no electron microscope or specimen holder modifications. Additionally, it eliminated the need to mount the sample onto specific patterned TEM grids or deposit markers, resulting in universal applications for most TEM samples, holders and electron microscopes for fast FOI identification. Furthermore, we implemented the method into a Gatan script for graphical-user-interface-based step-by-step instructions. Through online communication, the script enabled real-time navigation and tracking of the motion of samples in TEM on enlarged optical images with a panoramic view.

Image projective transformation rectification with synthetic data for smartphone-captured chest X-ray photos classification.

Automatic interpretation of chest X-ray (CXR) photos taken by smartphones at the same performance level as with digital CXRs is challenging, due to the projective transformation caused by the non-ideal camera position. Existing rectification methods for other camera-captured photos (document photos, license plate photos, etc.) cannot precisely rectify the projective transformation of CXR photos, due to its specific projective transformation type. In this paper, we propose an innovative deep learning-based Projective Transformation Rectification Network (PTRN) to automatically rectify the projective transformation of CXR photos by predicting the projective transformation matrix. Additionally, synthetic CXR photos are generated for training with the consideration of visual artifacts of natural images. The effectiveness of the proposed classification pipeline with PTRN is evaluated in the CheXphoto smartphone-captured CXR photo classification competition. It achieves first place with a huge performance improvement (ours 0.850, second-best 0.762, in AUC). Moreover, experimental results show that our approach successfully achieves the same performance level of digital CXR classification (AUC 0.893) on CXR photo classification (AUC 0.893).

Quantification of marine macro-debris abundance around Vancouver Island, Canada, based on archived aerial photographs processed by projective transformation.

The abundance of marine macro-debris was quantified with high spatial resolution by applying an image processing technique to archived shoreline aerial photographs taken over Vancouver Island, Canada. The photographs taken from an airplane at oblique angles were processed by projective transformation for georeferencing, where five reference points were defined by comparing aerial photographs with satellite images of Google Earth. Thereafter, pixels of marine debris were extracted based on their color differences from the background beaches. The debris abundance can be evaluated by the ratio of an area covered by marine debris to that of the beach (percent cover). The horizontal distribution of percent cover of marine debris was successfully computed from 167 aerial photographs and was significantly related to offshore Ekman flows and winds (leeway drift and Stokes drift). Therefore, the estimated percent cover is useful information to determine priority sites for mitigating adverse impacts across broad areas.