This project is an image classifier that detects and classifies images as either "happy" or "sad". It is built using TensorFlow and Keras and has been trained on a dataset containing images classified into these two categories.
Overview:
The goal of this project is to create a neural network that can accurately classify images as happy or sad. This can be useful for applications in emotion detection, sentiment analysis, and various AI-driven user experience improvements.
Loading Data:
We utilize the tf.keras.utils.image_dataset_from_directory API to handle data labeling, reshaping, and batching, making the data preprocessing pipeline straightforward and efficient.
Model Architecture:
The convolutional neural network is built using the Sequential API available in tensorflow.keras.models. The architecture includes the following layers: Conv2D: Convolutional layers for feature extraction MaxPooling2D: Max pooling layers for downsampling Flatten: Flattening the 2D arrays into a 1D vector Dense: Fully connected layers for classification
Result:
Training and Validation Accuracy:The model achieved an accuracy of approximately 100% on the training dataset, indicating that it learned the features and patterns associated with the "happy" and "sad" images effectively. During the validation phase, the model maintained a high accuracy, close to 100%, which suggests that the model generalized well to unseen data and did not overfit.
Loss and Accuracy Graphs:The training and validation loss graphs showed a steady decrease, and the accuracy graphs demonstrated an upward trend, which signifies that the model was learning effectively and improving with each epoch.
