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

A set of Deep Reinforcement Learning Agents implemented in Tensorflow.

A semismooth Newton method for elliptic PDE-constrained optimization

Send text to girlfriend in the morning

Code for the bachelors-thesis flaky fault localization

Fuzzer for Linux Kernel Drivers

ENet: A Deep Neural Network Architecture for Real-Time Semantic Segmentation

PyTorch implementation of paper "IBRNet: Learning Multi-View Image-Based Rendering", CVPR 2021.

improvement of CLIP features over the traditional resnet features on the visual question answering, image captioning, navigation and visual entailment tasks.

Source code of CIKM2021 Long Paper "PSSL: Self-supervised Learning for Personalized Search with Contrastive Sampling".

Evolutionary Population Curriculum for Scaling Multi-Agent Reinforcement Learning

Api's bulid in Flask perfom to manage Todo Task.

This thesis is mainly concerned with state-space methods for a class of deep Gaussian process (DGP) regression problems

Official pytorch implementation of "Feature Stylization and Domain-aware Contrastive Loss for Domain Generalization" ACMMM 2021 (Oral)

MemStream: Memory-Based Anomaly Detection in Multi-Aspect Streams with Concept Drift

Joint project of the duo Hacker Ninjas

Facial Image Inpainting with Semantic Control

Jittor implementation of PCT:Point Cloud Transformer

This repository gives an example on how to preprocess the data of the HECKTOR challenge

Dynamics-aware Adversarial Attack of 3D Sparse Convolution Network

Code release for "Masked-attention Mask Transformer for Universal Image Segmentation"