DeepLab and Bias Field Correction Based Automatic Cone Photoreceptor Cell Identification with Adaptive Optics Scanning Laser Ophthalmoscope Images

Yiwei Chen, Yi He, Jing Wang, Wanyue Li, Lina Xing, Xin Zhang, Guohua Shi

Abstract: The identification of cone photoreceptor cells is important for early diagnosing of eye diseases. We proposed automatic deep-learning cone photoreceptor cell identification on adaptive optics scanning laser ophthalmoscope images. The proposed algorithm is based on DeepLab and bias field correction. Considering manual identification as reference, our algorithm is highly effective, achieving precision, recall, and <svg xmlns:xlink="http://www.w3.org/1999/xlink" xmlns="http://www.w3.org/2000/svg" style="vertical-align:-0.0498209pt" id="M1" height="8.68572pt" version="1.1" viewBox="-0.0498162 -8.6359 14.2836 8.68572" width="14.2836pt"><g transform="matrix(.013,0,0,-0.013,0,0)"><path id="g113-71" d="M584 650H137L131 622C214 614 217 612 200 521L125 127C109 41 101 35 23 28L17 0H288L294 28C201 35 193 42 209 128L242 309H348C440 309 442 300 443 226H471L510 422H482C452 354 449 348 357 348H251L295 575C302 609 304 615 338 615H426C502 615 517 604 526 581C534 560 536 524 537 492L565 494C574 554 583 631 584 650Z"/></g><g transform="matrix(.013,0,0,-0.013,7.895,0)"><path id="g113-50" d="M384 0V27C293 34 287 42 287 114V635C232 613 172 594 109 583V559L157 557C201 555 205 550 205 499V114C205 42 199 34 109 27V0H384Z"/></g></svg> score of 96.7%, 94.6%, and 95.7%, respectively. To illustrate the performance of our algorithm, we present identification results for images with different cone photoreceptor cell distributions. The experimental results show that our algorithm can achieve accurate photoreceptor cell identification on images of human retinas, which is comparable to manual identification.