A video scene detection algorithm is designed to detect a variety of different scenes within a video

Last update: Jan 04, 2022

Overview

Scene-Change-Detection

The detection of scenes change is a simple problem that human beings face, but it gets much harder to handle autonomously a device that generally includes complex calculations and algorithms.

A video scene detection algorithm is designed to detect a variety of different scenes within a video. There is a very simple definition for a scene: It is a series of logically and chronologically related shots taken in a specific order to depict an over-arching concept or story. The identification of video scenes, in many video analysis applications, is a crucial pre-processing step. 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. Videos in the dataset have an open-source nature, which makes them an ideal product to be used by academics, as well as industry researchers alike.

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

A video scene detection algorithm is designed to detect a variety of different scenes within a video

Related tags

Overview

Scene-Change-Detection

DATASET

MODEL

LIBRARIES USED

COMPILATION USED

Training

REPORT

COLLABORATORS

License

Owner

Official implementation of paper Gradient Matching for Domain Generalization

NALSM: Neuron-Astrocyte Liquid State Machine

Use of Attention Gates in a Convolutional Neural Network / Medical Image Classification and Segmentation

Network Pruning That Matters: A Case Study on Retraining Variants (ICLR 2021)

A certifiable defense against adversarial examples by training neural networks to be provably robust

Dataset para entrenamiento de yoloV3 para 4 clases

Privacy as Code for DSAR Orchestration: Privacy Request automation to fulfill GDPR, CCPA, and LGPD data subject requests.

This repository contains all the code and materials distributed in the 2021 Q-Programming Summer of Qode.

A code implementation of AC-GC: Activation Compression with Guaranteed Convergence, in NeurIPS 2021.

This is a custom made virus code in python, using tkinter module.

Official code release for: EditGAN: High-Precision Semantic Image Editing

Towards Boosting the Accuracy of Non-Latin Scene Text Recognition

An efficient PyTorch implementation of the evaluation metrics in recommender systems.

End-to-End Object Detection with Fully Convolutional Network

Research code of ICCV 2021 paper "Mesh Graphormer"

This repo contains the code for paper Inverse Weighted Survival Games

DAT4 - General Assembly's Data Science course in Washington, DC

Open-source code for Generic Grouping Network (GGN, CVPR 2022)

All supplementary material used by me while TA-ing CS3244: Machine Learning

HIVE: Evaluating the Human Interpretability of Visual Explanations