Search and filter videos based on objects that appear in them using convolutional neural networks

Overview

Thingscoop: Utility for searching and filtering videos based on their content

Description

Thingscoop is a command-line utility for analyzing videos semantically - that means searching, filtering, and describing videos based on objects, places, and other things that appear in them.

When you first run thingscoop on a video file, it uses a convolutional neural network to create an "index" of what's contained in the every second of the input by repeatedly performing image classification on a frame-by-frame basis. Once an index for a video file has been created, you can search (i.e. get the start and end times of the regions in the video matching the query) and filter (i.e. create a supercut of the matching regions) the input using arbitrary queries. Thingscoop uses a very basic query language that lets you to compose queries that test for the presence or absence of labels with the logical operators ! (not), || (or) and && (and). For example, to search a video the presence of the sky and the absence of the ocean: thingscoop search 'sky && !ocean' <file>.

Right now two models are supported by thingscoop: vgg_imagenet uses the architecture described in "Very Deep Convolutional Networks for Large-Scale Image Recognition" to recognize objects from the ImageNet database, and googlenet_places uses the architecture described in "Going Deeper with Convolutions" to recognize settings and places from the MIT Places database. You can specify which model you'd like to use by running thingscoop models use <model>, where <model> is either vgg_imagenet or googlenet_places. More models will be added soon.

Thingscoop is based on Caffe, an open-source deep learning framework.

Installation

  1. Install ffmpeg, imagemagick, and ghostscript: brew install ffmpeg imagemagick ghostscript (Mac OS X) or apt-get install ffmpeg imagemagick ghostscript (Ubuntu).
  2. Follow the installation instructions on the Caffe Installation page.
  3. Make sure you build the Python bindings by running make pycaffe (on Caffe's directory).
  4. Set the environment variable CAFFE_ROOT to point to Caffe's directory: export CAFFE_ROOT=[Caffe's directory].
  5. Install thingscoop: easy_install thingscoop or pip install thingscoop.

Usage

thingscoop search <query> <files...>

Print the start and end times (in seconds) of the regions in <files> that match <query>. Creates an index for <file> using the current model if it does not exist.

Example output:

$ thingscoop search violin waking_life.mp4
/Users/anastasis/Downloads/waking_life.mp4 148.000000 162.000000
/Users/anastasis/Downloads/waking_life.mp4 176.000000 179.000000
/Users/anastasis/Downloads/waking_life.mp4 180.000000 186.000000
/Users/anastasis/Downloads/waking_life.mp4 189.000000 190.000000
/Users/anastasis/Downloads/waking_life.mp4 192.000000 200.000000
/Users/anastasis/Downloads/waking_life.mp4 211.000000 212.000000
/Users/anastasis/Downloads/waking_life.mp4 222.000000 223.000000
/Users/anastasis/Downloads/waking_life.mp4 235.000000 243.000000
/Users/anastasis/Downloads/waking_life.mp4 247.000000 249.000000
/Users/anastasis/Downloads/waking_life.mp4 251.000000 253.000000
/Users/anastasis/Downloads/waking_life.mp4 254.000000 258.000000

####thingscoop filter <query> <files...>

Generate a video compilation of the regions in the <files> that match <query>. Creates index for <file> using the current model if it does not exist.

Example output:

thingscoop sort <file>

Create a compilation video showing examples for every label recognized in the video (in alphabetic order). Creates an index for <file> using the current model if it does not exist.

Example output:

thingscoop describe <file>

Print every label that appears in <file> along with the number of times it appears. Creates an index for <file> using the current model if it does not exist.

thingscoop preview <file>

Create a window that plays the input video <file> while also displaying the labels the model recognizes on every frame.

$ thingscoop describe koyaanisqatsi.mp4 -m googlenet_places
sky 405
skyscraper 363
canyon 141
office_building 130
highway 78
lighthouse 66
hospital 64
desert 59
shower 49
volcano 45
underwater 44
airport_terminal 43
fountain 39
runway 36
assembly_line 35
aquarium 34
fire_escape 34
music_studio 32
bar 28
amusement_park 28
stage 26
wheat_field 25
butchers_shop 25
engine_room 24
slum 20
butte 20
igloo 20
...etc

thingscoop index <file>

Create an index for <file> using the current model if it does not exist.

thingscoop models list

List all models currently available in Thingscoop.

$ thingscoop models list
googlenet_imagenet            Model described in the paper "Going Deeper with Convolutions" trained on the ImageNet database
googlenet_places              Model described in the paper "Going Deeper with Convolutions" trained on the MIT Places database
vgg_imagenet                  16-layer model described in the paper "Return of the Devil in the Details: Delving Deep into Convolutional Nets" trained on the ImageNet database

thingscoop models info <model>

Print more detailed information about <model>.

$ thingscoop models info googlenet_places
Name: googlenet_places
Description: Model described in the paper "Going Deeper with Convolutions" trained on the MIT Places database
Dataset: MIT Places

thingscoop models freeze

List all models that have already been downloaded.

$ thingscoop models freeze
googlenet_places
vgg_imagenet

thingscoop models current

Print the model that is currently in use.

$ thingscoop models current
googlenet_places

thingscoop models use <model>

Set the current model to <model>. Downloads that model locally if it hasn't been downloaded already.

thingscoop models download <model>

Download the model <model> locally.

thingscoop models remove <model>

Remove the model <model> locally.

thingscoop models clear

Remove all models stored locally.

thingscoop labels list

Print all the labels used by the current model.

$ thingscoop labels list
abacus
abaya
abstraction
academic gown
accessory
accordion
acorn
acorn squash
acoustic guitar
act
actinic radiation
action
activity
adhesive bandage
adjudicator
administrative district
admiral
adornment
adventurer
advocate
...

thingscoop labels search <regexp>

Print all the labels supported by the current model that match the regular expression <regexp>.

$ thingscoop labels search instrument$
beating-reed instrument
bowed stringed instrument
double-reed instrument
free-reed instrument
instrument
keyboard instrument
measuring instrument
medical instrument
musical instrument
navigational instrument
negotiable instrument
optical instrument
percussion instrument
scientific instrument
stringed instrument
surveying instrument
wind instrument
...

Full usage options

thingscoop - Command-line utility for searching and filtering videos based on their content

Usage:
  thingscoop filter <query> <files>... [-o <output_path>] [-m <model>] [-s <sr>] [-c <mc>] [--recreate-index] [--gpu-mode] [--open]
  thingscoop search <query> <files>... [-o <output_path>] [-m <model>] [-s <sr>] [-c <mc>] [--recreate-index] [--gpu-mode] 
  thingscoop describe <file> [-n <words>] [-m <model>] [--recreate-index] [--gpu-mode] [-c <mc>]
  thingscoop index <files> [-m <model>] [-s <sr>] [-c <mc>] [-r <ocr>] [--recreate-index] [--gpu-mode] 
  thingscoop sort <file> [-m <model>] [--gpu-mode] [--min-confidence <ct>] [--max-section-length <ms>] [-i <ignore>] [--open]
  thingscoop preview <file> [-m <model>] [--gpu-mode] [--min-confidence <ct>]
  thingscoop labels list [-m <model>]
  thingscoop labels search <regexp> [-m <model>]
  thingscoop models list
  thingscoop models info <model>
  thingscoop models freeze
  thingscoop models current
  thingscoop models use <model>
  thingscoop models download <model>
  thingscoop models remove <model>
  thingscoop models clear

Options:
  --version                       Show version.
  -h --help                       Show this screen.
  -o --output <dst>               Output file for supercut
  -s --sample-rate <sr>           How many frames to classify per second (default = 1)
  -c --min-confidence <mc>        Minimum prediction confidence required to consider a label (default depends on model)
  -m --model <model>              Model to use (use 'thingscoop models list' to see all available models)
  -n --number-of-words <words>    Number of words to describe the video with (default = 5)
  -t --max-section-length <ms>    Max number of seconds to show examples of a label in the sorted video (default = 5)
  -r --min-occurrences <ocr>      Minimum number of occurrences of a label in video required for it to be shown in the sorted video (default = 2)
  -i --ignore-labels <labels>     Labels to ignore when creating the sorted video video
  --title <title>                 Title to show at the beginning of the video (sort mode only)
  --gpu-mode                      Enable GPU mode
  --recreate-index                Recreate object index for file if it already exists
  --open                          Open filtered video after creating it (OS X only)

CHANGELOG

0.2 (8/16/2015)

  • Added sort option for creating a video compilation of all labels appearing in a video
  • Now using JSON for the index files

0.1 (8/5/2015)

  • Conception

License

MIT

Owner
Anastasis Germanidis
🎭
Anastasis Germanidis
PyTorch implementations of the beta divergence loss.

Beta Divergence Loss - PyTorch Implementation This repository contains code for a PyTorch implementation of the beta divergence loss. Dependencies Thi

Billy Carson 7 Nov 09, 2022
Neural HMMs are all you need (for high-quality attention-free TTS)

Neural HMMs are all you need (for high-quality attention-free TTS) Shivam Mehta, Éva Székely, Jonas Beskow, and Gustav Eje Henter This is the official

Shivam Mehta 0 Oct 28, 2022
Revitalizing CNN Attention via Transformers in Self-Supervised Visual Representation Learning

Revitalizing CNN Attention via Transformers in Self-Supervised Visual Representation Learning This repository is the official implementation of CARE.

ChongjianGE 89 Dec 02, 2022
Repository for the NeurIPS 2021 paper: "Exploiting Domain-Specific Features to Enhance Domain Generalization".

meta-Domain Specific-Domain Invariant (mDSDI) Source code implementation for the paper: Manh-Ha Bui, Toan Tran, Anh Tuan Tran, Dinh Phung. "Exploiting

VinAI Research 12 Nov 25, 2022
Azua - build AI algorithms to aid efficient decision-making with minimum data requirements.

Project Azua 0. Overview Many modern AI algorithms are known to be data-hungry, whereas human decision-making is much more efficient. The human can re

Microsoft 197 Jan 06, 2023
This is an official implementation for "Swin Transformer: Hierarchical Vision Transformer using Shifted Windows" on Object Detection and Instance Segmentation.

Swin Transformer for Object Detection This repo contains the supported code and configuration files to reproduce object detection results of Swin Tran

Swin Transformer 1.4k Dec 30, 2022
Syed Waqas Zamir 906 Dec 30, 2022
A PyTorch implementation of the paper "Semantic Image Synthesis via Adversarial Learning" in ICCV 2017

Semantic Image Synthesis via Adversarial Learning This is a PyTorch implementation of the paper Semantic Image Synthesis via Adversarial Learning. Req

Seonghyeon Nam 146 Nov 25, 2022
Clustering with variational Bayes and population Monte Carlo

pypmc pypmc is a python package focusing on adaptive importance sampling. It can be used for integration and sampling from a user-defined target densi

45 Feb 06, 2022
Coded illumination for improved lensless imaging

CodedCam Coded Illumination for Improved Lensless Imaging Paper | Supplementary results | Data and Code are available. Coded illumination for improved

Computational Sensing and Information Processing Lab 1 Nov 29, 2021
NExT-QA: Next Phase of Question-Answering to Explaining Temporal Actions (CVPR2021)

NExT-QA We reproduce some SOTA VideoQA methods to provide benchmark results for our NExT-QA dataset accepted to CVPR2021 (with 1 'Strong Accept' and 2

Junbin Xiao 50 Nov 24, 2022
Weighted K Nearest Neighbors (kNN) algorithm implemented on python from scratch.

kNN_From_Scratch I implemented the k nearest neighbors (kNN) classification algorithm on python. This algorithm is used to predict the classes of new

1 Dec 14, 2021
Put blind watermark into a text with python

text_blind_watermark Put blind watermark into a text. Can be used in Wechat dingding ... How to Use install pip install text_blind_watermark Alice Pu

郭飞 164 Dec 30, 2022
GANsformer: Generative Adversarial Transformers Drew A

GANformer: Generative Adversarial Transformers Drew A. Hudson* & C. Lawrence Zitnick Update: We released the new GANformer2 paper! *I wish to thank Ch

Drew Arad Hudson 1.2k Jan 02, 2023
Code repository for the paper Computer Vision User Entity Behavior Analytics

Computer Vision User Entity Behavior Analytics Code repository for "Computer Vision User Entity Behavior Analytics" Code Description dataset.csv As di

Sameer Khanna 2 Aug 20, 2022
Riemann Noise Injection With PyTorch

Riemann Noise Injection - PyTorch A module for modeling GAN noise injection based on Riemann geometry, as described in Ruili Feng, Deli Zhao, and Zhen

2 May 27, 2022
Cave Generation using metaballs in Blender. Originally created by sdfgeoff, Edited by Myself (Archie Jaskowicz).

Blender-Cave-Generation Cave Generation using metaballs in Blender. Originally created by sdfgeoff, Edited by Myself (Archie Jaskowicz). Installation

2 Dec 28, 2022
G-NIA model from "Single Node Injection Attack against Graph Neural Networks" (CIKM 2021)

Single Node Injection Attack against Graph Neural Networks This repository is our Pytorch implementation of our paper: Single Node Injection Attack ag

Shuchang Tao 18 Nov 21, 2022
Json2Xml tool will help you convert from json COCO format to VOC xml format in Object Detection Problem.

JSON 2 XML All codes assume running from root directory. Please update the sys path at the beginning of the codes before running. Over View Json2Xml t

Nguyễn Trường Lâu 6 Aug 22, 2022
This is the implementation of our work Deep Extreme Cut (DEXTR), for object segmentation from extreme points.

This is the implementation of our work Deep Extreme Cut (DEXTR), for object segmentation from extreme points.

Sergi Caelles 828 Jan 05, 2023