Scene-Change-Detection

DATASET

• Dataset 2012 - In the dataset, there are six video categories, and in each category, there are four to six video sequences
• IBM video Scene Change Detection
• A dataset for video scene detection known as the Open Video Scene Detection (OVSD) dataset has been provided in order to evaluate algorithms for video scene detection.

MODEL

VGG16 was used for this project.VGG16 is a convolutional neural network model proposed by K. Simonyan and A. Zisserman from the University of Oxford in the paper “Very Deep Convolutional Networks for Large- Scale Image Recognition”. The model achieves 92.7% top-5 test accuracy in ImageNet, which is a dataset of over 14 million images belonging to 1000 classes.

LIBRARIES USED

• Numpy for array manipulation
• OpenCV (cv2) for Image Augmentation
• Keras for building the Neural Network
• Matplotlib for plotting visuals

COMPILATION USED

• Loss function selected is sparse categorical cross-entropy
• Optimizer selected is Adam
• Validation metric chosen is accuracy

Training

• No of epochs = 5
• Batch size = 1

REPORT

https://drive.google.com/file/d/1cwoP5cRJ5D76PvHV_WjCDRoSJIf0h9du/view?usp=sharing

COLLABORATORS

Neel kumar arya and Ashish Vidyarthi

License

MIT © Neel arya