Channel Randomisation with Domain Control for Effective Representation Learning of Visual Anomalies in Strawberries

Choi, Taeyeong and Cielniak, Grzegorz (2022) Channel Randomisation with Domain Control for Effective Representation Learning of Visual Anomalies in Strawberries. In: AI for Agriculture and Food Systems, 28-2-2022, Virtual.

Documents
Channel Randomisation with Domain Control for Effective Representation Learning of Visual Anomalies in Strawberries

Request a copy
[img] PDF
channel_randomisation_with_dom.pdf - Whole Document
Restricted to Repository staff only

38MB
Item Type:Conference or Workshop contribution (Presentation)
Item Status:Live Archive

Abstract

Channel Randomisation (CH-Rand) has appeared as a key data augmentation technique for anomaly detection on fruit
images because neural networks can learn useful representations of colour irregularity whilst classifying the samples
from the augmented "domain". Our previous study has revealed its success with significantly more reliable performance than other state-of-the-art methods, largely specialised for identifying structural implausibility on non-agricultural objects (e.g., screws). In this paper, we further enhance CH-Rand with additional guidance to generate more informative data for representation learning of anomalies in fruits as most of its fundamental designs are still maintained. To be specific, we first control the "colour space" on which CH-Rand is executed to investigate whether a particular model—e.g., HSV , YCbCr, or L*a*b* —can better help synthesise realistic anomalies than the RGB, suggested in the original design. In addition, we develop a learning "curriculum" in which CH-Rand shifts its augmented domain to gradually increase the difficulty of the examples for neural networks to classify. To the best of our best knowledge, we are the first to connect the concept of curriculum to self-supervised representation learning for anomaly detection. Lastly, we perform evaluations with the Riseholme-2021 dataset, which contains > 3.5K real strawberry images at various growth levels along with anomalous examples. Our experimental results show that the trained models with the proposed strategies can achieve over 16% higher scores of AUC-PR with more than three times less variability than the naive CH-Rand whilst using the same deep networks and data.

Keywords:Self-supervised Learning, Anomaly Detection, Environmental monitoring, Artificial Intelligence, Machine Learning, one-class classifier, Computer Vision, Mobile robotics, Robot vision systems, Agricultural Robotics, Deep Neural Networks, Convolutional Neural Networks, Curriculum Learning, Fruit Image Data
Subjects:G Mathematical and Computer Sciences > G700 Artificial Intelligence
G Mathematical and Computer Sciences > G400 Computer Science
G Mathematical and Computer Sciences > G760 Machine Learning
H Engineering > H671 Robotics
G Mathematical and Computer Sciences > G740 Computer Vision
D Veterinary Sciences, Agriculture and related subjects > D400 Agriculture
Divisions:College of Science > Lincoln Institute for Agri-Food Technology
ID Code:48676
Deposited On:31 Mar 2022 10:15

Repository Staff Only: item control page