Ensemble of top3 prediction with image pixel interval method using deep learning

Computer vision (CV) has been successfully used in picture categorization applications various fields, including medicine, production quality control, and transportation systems. CV models use an excessive number of photos to train potential models. Considering that image acquisition is typically expensive time-consuming, this study, we provide a multistep strategy improve accuracy with less data. In the first stage, constructed numerous datasets from single dataset. Given pixels values ranging 0 255, images were separated into pixel intervals based on type The interval was split two portions when dataset grayscale five it composed RGB images. Next, trained model using both original newly datasets. Each training process showed non-identical prediction space, suggested top three probability ensemble technique. predictions for created combined corresponding image. results learning patterns each ensembling significantly performance accuracy, can be any model.