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

Pytorch implementation of CVPR2021 paper "MUST-GAN: Multi-level Statistics Transfer for Self-driven Person Image Generation"

darija <-> english dictionary

Revisiting Discriminator in GAN Compression: A Generator-discriminator Cooperative Compression Scheme (NeurIPS2021)

A Shading-Guided Generative Implicit Model for Shape-Accurate 3D-Aware Image Synthesis

Learning Off-Policy with Online Planning, CoRL 2021

Code release for NeRF (Neural Radiance Fields)

An efficient PyTorch library for Global Wheat Detection using YOLOv5. The project is based on this Kaggle competition Global Wheat Detection (2021).

code for ICCV 2021 paper 'Generalized Source-free Domain Adaptation'

Official implementation for: Blended Diffusion for Text-driven Editing of Natural Images.

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

OpenFace – a state-of-the art tool intended for facial landmark detection, head pose estimation, facial action unit recognition, and eye-gaze estimation.

TLoL (Python Module) - League of Legends Deep Learning AI (Research and Development)

AISTATS 2019: Confidence-based Graph Convolutional Networks for Semi-Supervised Learning

Official PyTorch implementation of SyntaSpeech (IJCAI 2022)

A LiDAR point cloud cluster for panoptic segmentation

Self-supervised learning optimally robust representations for domain generalization.

ESTDepth: Multi-view Depth Estimation using Epipolar Spatio-Temporal Networks (CVPR 2021)

🥇 LG-AI-Challenge 2022 1위 솔루션 입니다.

Exploit Camera Raw Data for Video Super-Resolution via Hidden Markov Model Inference

The aim of the game, as in the original one, is to find a specific image from a group of different images of a person's face