Improving Localization for Semi-Supervised Object Detection

TL;DR

This paper enhances semi-supervised object detection by adding a localization classification task, improving pseudo-label filtering, and demonstrating that bounding box regression benefits training, leading to a 1.14% AP increase on COCO.

Contribution

Introduces IL-net, a method that improves localization in SSOD by adding a localization classification task and validating the role of bounding box regression in training.

Findings

IL-net increases SSOD performance by 1.14% AP on COCO.

Bounding box regression contributes equally to training as category classification.

Enhanced filtering of pseudo-labels improves detection quality.

Abstract

Nowadays, Semi-Supervised Object Detection (SSOD) is a hot topic, since, while it is rather easy to collect images for creating a new dataset, labeling them is still an expensive and time-consuming task. One of the successful methods to take advantage of raw images on a Semi-Supervised Learning (SSL) setting is the Mean Teacher technique, where the operations of pseudo-labeling by the Teacher and the Knowledge Transfer from the Student to the Teacher take place simultaneously. However, the pseudo-labeling by thresholding is not the best solution since the confidence value is not strictly related to the prediction uncertainty, not permitting to safely filter predictions. In this paper, we introduce an additional classification task for bounding box localization to improve the filtering of the predicted bounding boxes and obtain higher quality on Student training. Furthermore, we empirically prove that bounding box regression on the unsupervised part can equally contribute to the training as much as category classification. Our experiments show that our IL-net (Improving Localization net) increases SSOD performance by 1.14% AP on COCO dataset in limited-annotation regime. The code is available at https://github.com/IMPLabUniPr/unbiased-teacher/tree/ilnet