Building extraction from aerial and satellite imagery is a crucial task for applications such as urban planning, disaster management, and geospatial analysis. Despite recent advances in deep learning, the automatic extraction of buildings remains challenging due to the diversity of building characteristics and surrounding complexities. In this work, we explore the effectiveness of a single-stage instance segmentation model for building extraction as a potentially better alternative to two-stage models. Namely, we employ RTMDet, a state-of-the-art single-stage instance segmentation model, to investigate its performance in building extraction tasks. RTMDet offers flexibility between segmentation performance and inference speed, which we assess in terms of accuracy and efficiency. Furthermore, we examine the impact of data augmentation techniques, particularly affine transformations, and the impact of the amount of training data on improving model performance. Experimental results demonstrate a comprehensive comparison of several training configurations as well as a comparison between single-stage and two-stage instance segmentation approaches. Our source codes and the best-performing model parameters are available at https://github.com/tamtamz/bigdata2024-buildext.

We propose a simple, intuitive yet powerful method for human-object interaction (HOI) detection. HOIs are so diverse in spatial distribution in an image that existing CNN-based methods face the following three major drawbacks; they cannot leverage image-wide features due to CNN’s locality, they rely on a manually defined location-of-interest for the feature aggregation, which sometimes does not cover contextually important regions, and they cannot help but mix up the features for multiple HOI instances if they are located closely. To overcome these drawbacks, we propose a transformer-based feature extractor, in which an attention mechanism and query-based detection play key roles. The attention mechanism is effective in aggregating contextually important information imagewide, while the queries, which we design in such a way that each query captures at most one human-object pair, can avoid mixing up the features from multiple instances. This transformer-based feature extractor produces so effective embeddings that the subsequent detection heads may be fairly simple and intuitive. The extensive analysis reveals that the proposed method successfully extracts contextually important features, and thus outperforms existing methods by large margins (5.37 mAP on HICO-DET, and 5.6 mAP on V-COCO). The source codes are available at https://github.com/hitachi-rd-cv/qpic.

Recently much progress has been made in pedestrian detection by utilizing the learning ability of convolutional neural networks (CNNs). However, due to the lack of omnidirectional images to train CNNs, few CNN-based detectors have been proposed for omnidirectional pedestrian detection. One significant difference between omnidirectional images and perspective images is that the appearance of pedestrians is rotated in omnidirectional images. A previous method has dealt with this by transforming omnidirectional images into perspective images in the test phase. However, this method has significant drawbacks, namely, the computational cost and the performance degradation caused by the transformation. To address this issue, we propose a rotation invariant training method, which only uses randomly rotated perspective images without any additional annotation. By this method, existing large-scale datasets can be utilized. In test phase, omnidirectional images can be used without the transformation. To group predicted bounding boxes, we also develop a bounding box refinement, which works better for our detector than non-maximum suppression. The proposed detector achieved a state-of-the-art performance on four public benchmarks.