Towards on-farm pig face recognition using convolutional neural networks

MF Hansen, Melvyn L Smith, LN Smith, MG Salter, EM Baxter, M Farish, B Grieve

Research output: Contribution to journalArticle

18 Citations (Scopus)
13 Downloads (Pure)

Abstract

Identification of individual livestock such as pigs and cows has become a pressing issue in recent years as intensification practices continue to be adopted and precise objective measurements are required (e.g. weight). Current best practice involves the use of RFID tags which are time-consuming for the farmer and distressing for the animal to fit. To overcome this, non-invasive biometrics are proposed by using the face of the animal. We test this in a farm environment, on 10 individual pigs using three techniques adopted from the human face recognition literature: Fisherfaces, the VGG-Face pre-trained face convolutional neural network (CNN) model and our own CNN model that we train using an artificially augmented data set. Our results show that accurate individual pig recognition is possible with accuracy rates of 96.7% on 1553 images. Class Activated Mapping using Grad-CAM is used to show the regions that our network uses to discriminate between pigs.
Original languageEnglish
Pages (from-to)145 - 152
Number of pages8
JournalComputers in Industry
Volume98
Early online date21 Mar 2018
DOIs
Publication statusFirst published - 21 Mar 2018

Fingerprint

Face recognition
Farms
Animals
Neural networks
Computer aided manufacturing
Biometrics
Radio frequency identification (RFID)

Bibliographical note

1031029

Keywords

  • Biometrics
  • Convolutional neural network
  • Deep learning
  • Pig face recognition

Cite this

Hansen, MF., Smith, M. L., Smith, LN., Salter, MG., Baxter, EM., Farish, M., & Grieve, B. (2018). Towards on-farm pig face recognition using convolutional neural networks. Computers in Industry, 98, 145 - 152. https://doi.org/10.1016/j.compind.2018.02.016
Hansen, MF ; Smith, Melvyn L ; Smith, LN ; Salter, MG ; Baxter, EM ; Farish, M ; Grieve, B. / Towards on-farm pig face recognition using convolutional neural networks. In: Computers in Industry. 2018 ; Vol. 98. pp. 145 - 152.
@article{18b7ecf5ec0b4d93a1e24436ebeb60d6,
title = "Towards on-farm pig face recognition using convolutional neural networks",
abstract = "Identification of individual livestock such as pigs and cows has become a pressing issue in recent years as intensification practices continue to be adopted and precise objective measurements are required (e.g. weight). Current best practice involves the use of RFID tags which are time-consuming for the farmer and distressing for the animal to fit. To overcome this, non-invasive biometrics are proposed by using the face of the animal. We test this in a farm environment, on 10 individual pigs using three techniques adopted from the human face recognition literature: Fisherfaces, the VGG-Face pre-trained face convolutional neural network (CNN) model and our own CNN model that we train using an artificially augmented data set. Our results show that accurate individual pig recognition is possible with accuracy rates of 96.7{\%} on 1553 images. Class Activated Mapping using Grad-CAM is used to show the regions that our network uses to discriminate between pigs.",
keywords = "Biometrics, Convolutional neural network, Deep learning, Pig face recognition",
author = "MF Hansen and Smith, {Melvyn L} and LN Smith and MG Salter and EM Baxter and M Farish and B Grieve",
note = "1031029",
year = "2018",
month = "3",
day = "21",
doi = "10.1016/j.compind.2018.02.016",
language = "English",
volume = "98",
pages = "145 -- 152",
journal = "Computers in Industry",
issn = "0166-3615",
publisher = "Elsevier",

}

Hansen, MF, Smith, ML, Smith, LN, Salter, MG, Baxter, EM, Farish, M & Grieve, B 2018, 'Towards on-farm pig face recognition using convolutional neural networks', Computers in Industry, vol. 98, pp. 145 - 152. https://doi.org/10.1016/j.compind.2018.02.016

Towards on-farm pig face recognition using convolutional neural networks. / Hansen, MF; Smith, Melvyn L; Smith, LN; Salter, MG; Baxter, EM; Farish, M; Grieve, B.

In: Computers in Industry, Vol. 98, 21.03.2018, p. 145 - 152.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Towards on-farm pig face recognition using convolutional neural networks

AU - Hansen, MF

AU - Smith, Melvyn L

AU - Smith, LN

AU - Salter, MG

AU - Baxter, EM

AU - Farish, M

AU - Grieve, B

N1 - 1031029

PY - 2018/3/21

Y1 - 2018/3/21

N2 - Identification of individual livestock such as pigs and cows has become a pressing issue in recent years as intensification practices continue to be adopted and precise objective measurements are required (e.g. weight). Current best practice involves the use of RFID tags which are time-consuming for the farmer and distressing for the animal to fit. To overcome this, non-invasive biometrics are proposed by using the face of the animal. We test this in a farm environment, on 10 individual pigs using three techniques adopted from the human face recognition literature: Fisherfaces, the VGG-Face pre-trained face convolutional neural network (CNN) model and our own CNN model that we train using an artificially augmented data set. Our results show that accurate individual pig recognition is possible with accuracy rates of 96.7% on 1553 images. Class Activated Mapping using Grad-CAM is used to show the regions that our network uses to discriminate between pigs.

AB - Identification of individual livestock such as pigs and cows has become a pressing issue in recent years as intensification practices continue to be adopted and precise objective measurements are required (e.g. weight). Current best practice involves the use of RFID tags which are time-consuming for the farmer and distressing for the animal to fit. To overcome this, non-invasive biometrics are proposed by using the face of the animal. We test this in a farm environment, on 10 individual pigs using three techniques adopted from the human face recognition literature: Fisherfaces, the VGG-Face pre-trained face convolutional neural network (CNN) model and our own CNN model that we train using an artificially augmented data set. Our results show that accurate individual pig recognition is possible with accuracy rates of 96.7% on 1553 images. Class Activated Mapping using Grad-CAM is used to show the regions that our network uses to discriminate between pigs.

KW - Biometrics

KW - Convolutional neural network

KW - Deep learning

KW - Pig face recognition

U2 - 10.1016/j.compind.2018.02.016

DO - 10.1016/j.compind.2018.02.016

M3 - Article

VL - 98

SP - 145

EP - 152

JO - Computers in Industry

JF - Computers in Industry

SN - 0166-3615

ER -