“…We get ξ disc (D * ) and ξ mac (M * ) values of 3.74% and 3.94%, which is considerably lower than 7% reported in [9]. On STARE dataset alone, the overall sensitivity of OD detection is 95%, comparable to reported results of [8], whereas in macula detection, our method achieves sensitivity 97.53%, an improvement over the best published results of [6]. Based on our extensive evaluation and results of the intermediate stages of our integration method, we tabulate the error percentages observed among true-positives in each dataset of our collection, along with Combined (shown in On pathological datasets such as DMED (macular edema), DIARETDB1, MESSIDOR (DR), our method shows low error for M * with nil false-positives, indicating its potential in disease grading applications.…”