CIFAR10 Classification Model

A classification model implemented using Deep Neural Networks

Algorithms Used

Convolutional neural network

Popularly used for Image Processing, CNN's are a series of convolutional, nonlinear, pooling (downsampling), and fully connected layers that produces an output. This output can be a single class or a probability of classes that best describes the image. Convolutions use a kernel matrix to scan a given image and apply a filter to obtain a certain effect and still maintains the spatial relationship between pixels.

Confusion Matrix

• A confusion matrix is used to describe the performance of a classification model:
• True positives (TP): Classifier predicted TRUE and correct class was TRUE.
• True negatives (TN): Classifier predicted FALSE and correct class was FALSE .
• False positives (FP): Classifier predicted TRUE but correct class was FALSE.
• False negatives (FN): classifier predicted FALSE but correct class was TRUE.

Implementation

##Dataset CIFAR-10 is a dataset that consists of several images divided into the following 10 classes:

• Airplanes
• Cars
• Bird
• Cats
• Deer
• Dogs
• Frogs
• Horses
• Ships
• Trucks It consists of 60,000 images with low resolution (32x32). After the model is trained, it classifies the testing dataset to one of the category as shown above.