Internet of Things (IoT) and cloud computing are two related domains that depends on each other through which the physicians monitor and support the remote patients. Effective treatment of diseases need a model designed with IoT equipments to identify the diseases. This paper proposes a novel automated glaucoma recognition method for identifying the lesion region present in the given fundus image. The proposed BFOA-CNN model executes in five steps, viz., i) calibrate the image, ii) preprocess the image, iii) segment the image, iv) extract the features and v) classify the image. The image given as input is first calibrated and involved into preprocessing stage so as to make the image suitable to process further for successful classification. Next, optic disc is removed by the entropy method and CRF (Conditional Random Fields) based segmentation process is used for detecting the presence of lesion region in the given fundus image. Further the extraction of shape features is executed and it is followed by the classification using Random Forest classification model. BFOA-CNN model performance is investigated with a set of 101 retinal fundus image and the results proved the best efficiency of the model. Accuracy: Internet of Things, BFOA-CNN, Retinal fundus image, Glaucoma Recognition, Conditional Random Fields (CRF).