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

The Self-Supervised Learner can be used to train a classifier with fewer labeled examples needed using self-supervised learning.

DEMix Layers for Modular Language Modeling

Official implementation of Monocular Quasi-Dense 3D Object Tracking

Patches desktop steam to look like the new steamdeck ui.

2020 CCF大数据与计算智能大赛-非结构化商业文本信息中隐私信息识别-第7名方案

[TPAMI 2021] iOD: Incremental Object Detection via Meta-Learning

Implementation of "Fast and Flexible Temporal Point Processes with Triangular Maps" (Oral @ NeurIPS 2020)

TLXZoo - Pre-trained models based on TensorLayerX

Point-NeRF: Point-based Neural Radiance Fields

Classifying audio using Wavelet transform and deep learning

Download and preprocess popular sequential recommendation datasets

Collect some papers about transformer with vision. Awesome Transformer with Computer Vision (CV)

Lazy, a tool for running things in idle time

Ego4d dataset repository. Download the dataset, visualize, extract features & example usage of the dataset

Demo notebooks for Qiskit application modules demo sessions (Oct 8 & 15):

A curated list of awesome open source libraries to deploy, monitor, version and scale your machine learning

A simple pytorch pipeline for semantic segmentation.

Chainer Implementation of Semantic Segmentation using Adversarial Networks

Demo for the paper "Overlap-aware low-latency online speaker diarization based on end-to-end local segmentation"

A framework for the elicitation, specification, formalization and understanding of requirements.