Object Detection

Use the CNN model to classify objects

CIFAR10 data set was used in this project. Object detection was performed on this data set using CNN. The aim of the project is to observe the effect of the parameters and methods used in the model on the model accuracy score while object detection. Therefore, various parameters and methods were used in the CNN models created in the project, six models were created to observe the performances. The results of the models are shown as plots and table. Thus, all models are compared.

Import the required packages

• numpy
• pandas
• seaborn
• matplotlib
• warnings
• keras
• tensorflow

Content

• Preprocess
• Model building without augmentation, without batch normalization and without dropout.
• Model building without dropout, without batch normalization and with augmentation.
• Model building without, with batch normalization and with augmentation.
• Model building with dropout, batch normalization and augmentation.
• Model building less pooling layer.
• Model building extra hidden layer.
• Results

Data Set

For the data set in the project, https://www.kaggle.com/ was used. Data set link used:

https://keras.io/api/datasets/cifar10/