DATA COLLECTION AND LABELLING

PRE PROCESSING AND AUGUMENTATION

FEATURE EXTRACTION IN .H5 FILE

MODELS TRAINED TO GET THE BEST MODEL

ViT-G/14

DaViT-Giant

CoCa

BUILD ,TRAIN COMPILE,EVALUATE THE HYBRID MODEL

BUILD ,TRAIN COMPILE,EVALUATE THE HYBRID MODEL

BUILD ,TRAIN COMPILE,EVALUATE THE HYBRID MODEL

1.CLASSIFICATION TAKES PLACE 2.GRADCAM FOR WRONG PREDICTION-WITH ACCURACY

1.CLASSIFICATION TAKES PLACE 2.GRADCAM FOR WRONG PREDICTION-WITH ACCURACY

1.CLASSIFICATION TAKES PLACE 2.GRADCAM FOR WRONG PREDICTION-WITH ACCURACY

EVALUATE BEST MODEL ON THE BASIS OF ACCURACY (referring phase 1 models)

CAPTIONING TAKES PLACE BASED ON THE CLASSIFICATION’S SYMPTOMS.

WEB APPLICATION