Low-Cost, Continuous Motion Imaging, Computationally Augmented Whole Slide Imager for Digital Pathology

As digital pathology advances, systems for acquiring Whole Slide Images (WSI) need to become affordable and effective. Current research is majorly reported around the low-throughput stop-and-focus imaging systems that are too slow for clinical use. This work reports a system that acquires well focused images at a high throughput. It reduces the imaging time by an order of magnitude via the carefully engineered implementation of a global shutter camera which allows continuous motion imaging. The second challenge of maintaining good focus comes from three sources: the imperfections induced by low-cost optomechanics, smear surface undulations, and irregularity of the slide surface itself. These are minimized via a tiled-scan method supported by focal mapping, which predicts and achieves best focus on-the-fly. Finally, any remnant defocus is compensated for by a post-acquisition image enhancement method which sharpens the image without disturbing the essential features and color tones. Thus, this system delivers a 15 × 15 mm ² scan in about 3 minutes at a resolution of 0.78 μm with a 40x microscopy objective. The bill of material cost is about US $1300 and hence it would be beneficial for telepathology in resource-constrained scenarios. Also, this device can serve as an enabler of Artificial Intelligence/Machine Learning algorithms to provide fully automated diagnosis. The developed codes for post-acquisition image enhancement are available at https://github.com/navchetan-awasthi/Microscopy.