TL;DR: We present BALF, a simple yet both efficient and effective motion blur aware local feature detector.
Local feature detection is a key ingredient of many image processing and computer vision applications, such as visual odometry and localization. Most existing algorithms focus on feature detection from a sharp image. They would thus have degraded performance once the image is blurred, which could happen easily under low-lighting conditions. To address this issue, we propose a simple yet both efficient and effective keypoint detection method that is able to accurately localize the salient keypoints in a blurred image. Our method takes advantages of a novel multi-layer perceptron (MLP) based architecture that significantly improve the detection repeatability for a blurred image. The network is also light-weight and able to run in real-time, which enables its deployment for time-constrained applications. Extensive experimental results demonstrate that our detector is able to improve the detection repeatability with blurred images, while keeping comparable performance as existing state-of-the-art detectors for sharp images.
The proposed network for motion blur aware local feature detector (BALF). Our network consists of two main modules: a multi-stage MLP-based encoder to extract an intermediate feature representation of the input image, and a MLP detection module to detect salient keypoints via a differentiable softmax operator.
Detection module. The learned feature representations of the input image is processed by two channel-wise MLP blocks. The keypoints are then detected by using channel-wise softmax operation and mapped back to the original image domain.
Our BALF achieves superior detection performance over prior works on motion blurred images, while keeping comparable performance for sharp images.
Single image deblurring network can indeed help with the local feature detection. However, it still cannot perform on-par with our one-stage detection network without doing any intermediate deblurring operation.
Our BALF is able to run in real-time (&tilde34.46 FPS) with a VGA resolution image (i.e. 480×640 pixels).
Our BALF is able to generate more accurate and consistent keypoints in a blurred image.
Our BALF is able to detect well distributed and localized keypoints from both sharp and blurred images for further image matching.
@InProceedings{Zhao_2024_WACV,
author = {Zhao, Zhenjun},
title = {BALF: Simple and Efficient Blur Aware Local Feature Detector},
booktitle = {Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV)},
month = {January},
year = {2024},
pages = {3362-3372}
}