Transformation from hematoxylin-and-eosin staining to Ki-67 immunohistochemistry digital staining images using deep learning: experimental validation on the labeling index.

Purpose: Endometrial cancer (EC) is one of the most common types of cancer affecting women. While the hematoxylin-and-eosin (H&E) staining remains the standard for histological analysis, the immunohistochemistry (IHC) method provides molecular-level visualizations. Our study proposes a digital staining method to generate the hematoxylin-3,3'-diaminobenzidine (H-DAB) IHC stain of Ki-67 for the whole slide image of the EC tumor from its H&E stain counterpart. Approach: We employed a color unmixing technique to yield stain density maps from the optical density (OD) of the stains and utilized the U-Net for end-to-end inference. The effectiveness of the proposed method was evaluated using the Pearson correlation between the digital and physical stain's labeling index (LI), a key metric indicating tumor proliferation. Two different cross-validation schemes were designed in our study: intraslide validation and cross-case validation (CCV). In the widely used intraslide scheme, the training and validation sets might include different regions from the same slide. The rigorous CCV validation scheme strictly prohibited any validation slide from contributing to training. Results: The proposed method yielded a high-resolution digital stain with preserved histological features, indicating a reliable correlation with the physical stain in terms of the Ki-67 LI. In the intraslide scheme, using intraslide patches resulted in a biased accuracy (e.g., R = 0.98 ) significantly higher than that of CCV. The CCV scheme retained a fair correlation (e.g., R = 0.66 ) between the LIs calculated from the digital stain and its physical IHC counterpart. Inferring the OD of the IHC stain from that of the H&E stain enhanced the correlation metric, outperforming that of the baseline model using the RGB space. Conclusions: Our study revealed that molecule-level insights could be obtained from H&E images using deep learning. Furthermore, the improvement brought via OD inference indicated a possible method for creating more generalizable models for digital staining via per-stain analysis.

Wave-optics-based image synthesis for super resolution reconstruction of a FZA lensless camera.

A Fresnel Zone Aperture (FZA) mask for a lensless camera, an ultra-thin and functional computational imaging system, is beneficial because the FZA pattern makes it easy to model the imaging process and reconstruct captured images through a simple and fast deconvolution. However, diffraction causes a mismatch between the forward model used in the reconstruction and the actual imaging process, which affects the recovered image's resolution. This work theoretically analyzes the wave-optics imaging model of an FZA lensless camera and focuses on the zero points caused by diffraction in the frequency response. We propose a novel idea of image synthesis to compensate for the zero points through two different realizations based on the linear least-mean-square-error (LMSE) estimation. Results from computer simulation and optical experiments verify a nearly two-fold improvement in spatial resolution from the proposed methods compared with the conventional geometrical-optics-based method.

Computational see-through screen camera based on a holographic waveguide device.

This study proposes a novel computational imaging system that integrates a see-through screen (STS) with volume holographic optical elements (vHOEs) and a digital camera unit. Because of the unique features of the vHOE, the STS can function as a holographic waveguide device (HWD) and enable the camera to capture the frontal image when the user gazes at the screen. This system not only provides an innovative solution to a high-quality video communication system by realizing eye-contact but also contributes to other visual applications due to its refined structure. However, there is a dilemma in the proposed imaging system: for a wider field of view, a larger vHOE is necessary. If the size of the vHOE is larger, the light rays from the same object point are diffracted at different Bragg conditions and reflect a different number of times, which causes blurring of the captured image. The system imaging process is analyzed by ray tracing, and a digital image reconstruction method was employed to obtain a clear picture in this study. Optical experiments confirmed the effectiveness of the proposed HWD-STS camera.

Image reconstruction with transformer for mask-based lensless imaging.

A mask-based lensless camera optically encodes the scene with a thin mask and reconstructs the image afterward. The improvement of image reconstruction is one of the most important subjects in lensless imaging. Conventional model-based reconstruction approaches, which leverage knowledge of the physical system, are susceptible to imperfect system modeling. Reconstruction with a pure data-driven deep neural network (DNN) avoids this limitation, thereby having potential to provide a better reconstruction quality. However, existing pure DNN reconstruction approaches for lensless imaging do not provide a better result than model-based approaches. We reveal that the multiplexing property in lensless optics makes global features essential in understanding the optically encoded pattern. Additionally, all existing DNN reconstruction approaches apply fully convolutional networks (FCNs) which are not efficient in global feature reasoning. With this analysis, for the first time to the best of our knowledge, a fully connected neural network with a transformer for image reconstruction is proposed. The proposed architecture is better in global feature reasoning, and hence enhances the reconstruction. The superiority of the proposed architecture is verified by comparing with the model-based and FCN-based approaches in an optical experiment.