Combining Supervised and Un-supervised Learning for Automatic Citrus Segmentation

TL;DR

This paper introduces a hybrid approach combining supervised CNN and unsupervised movement learning via multimodal transformers to improve citrus segmentation accuracy with less labeled data and temporal information.

Contribution

It proposes a novel combination of supervised and unsupervised learning using multimodal transformers for citrus segmentation, leveraging limited labeled data and unlabeled videos.

Findings

Achieved 88.3% IOU and 93.6% precision in citrus segmentation.

Outperformed baseline methods with 1.2% higher IOU and 2.4% higher precision.

Effectively utilized small labeled datasets and large unlabeled video data.

Abstract

Citrus segmentation is a key step of automatic citrus picking. While most current image segmentation approaches achieve good segmentation results by pixel-wise segmentation, these supervised learning-based methods require a large amount of annotated data, and do not consider the continuous temporal changes of citrus position in real-world applications. In this paper, we first train a simple CNN with a small number of labelled citrus images in a supervised manner, which can roughly predict the citrus location from each frame. Then, we extend a state-of-the-art unsupervised learning approach to pre-learn the citrus's potential movements between frames from unlabelled citrus's videos. To take advantages of both networks, we employ the multimodal transformer to combine supervised learned static information and unsupervised learned movement information. The experimental results show that combing both network allows the prediction accuracy reached at 88.3$\%$ IOU and 93.6$\%$ precision, outperforming the original supervised baseline 1.2$\%$ and 2.4$\%$. Compared with most of the existing citrus segmentation methods, our method uses a small amount of supervised data and a large number of unsupervised data, while learning the pixel level location information and the temporal information of citrus changes to enhance the segmentation effect.