Using Weakly Supervised Semantic Segmentation for Post-Disaster Scene Understanding
Publication: Construction Research Congress 2024
ABSTRACT
Creating semantic segmentation datasets is more time-consuming than the preparation of image-level classification labels. This paper presents a cost-efficient pixel labeling method and experimental results of using Grad-CAM-generated heatmap for weakly supervised post-disaster site semantic segmentation. It starts with training CNN binary-classification models, then removing non-target objects, reducing searching regions according to the distributions of heatmap values, and assigning the pixel-level label to the highest heatmap value regions as segmentations. The proposed CNN model architecture has a 1 × 1 convolutional heatmap layer with a suitable number of channels to fit the target objects. Testing results showed the 16-channel heatmap layer reached the best building-background classification performance (F1-score 0.89), and the 128-channel had the best building-damage classification performance (F1-score 0.78) among the evaluated options of 4, 8, 16, 32, 64, and 128 channels. Furthermore, the proposed pixel labeling method can understand complicated post-disaster scenes by correctly locating target objects (buildings and damages) in aerial images via repetitions of searching the target objects in regions of higher heatmap values.
Get full access to this article
View all available purchase options and get full access to this chapter.
REFERENCES
Chattopadhay, A., Sarkar, A., Howlader, P., and Balasubramanian, V. N. (2018). “Grad-CAM++: Generalized Gradient-Based Visual Explanations for Deep Convolutional Networks.” 2018 WACV, IEEE, 839–847.
Han, S., Chung, I., Jiang, Y., and Uwakweh, B. (2023). “PCIer: Pavement Condition Evaluation Using Aerial Imagery and Deep Learning.” Geographies, 3(1), 132–142.
keras.io. (2021). “VGG16 and VGG19.” <https://keras.io/api/applications/vgg/>(Feb. 27, 2021).
Li, Q., Arnab, A., and Torr, P. H. S. (2018). “Weakly- and Semi-supervised Panoptic Segmentation.” Lecture Notes in Computer Science, 106–124.
Oh, S. J., Benenson, R., Khoreva, A., Akata, Z., Fritz, M., and Schiele, B. (2017). “Exploiting Saliency for Object Segmentation from Image Level Labels.” 2017 CVPR, IEEE, 5038–5047.
OpenCV. (2020). “Image Segmentation with Watershed Algorithm.” <https://docs.opencv.org/3.4.12/d3/db4/tutorial_py_watershed.html>(Mar. 26, 2021).
Rahnemoonfar, M., Chowdhury, T., Sarkar, A., Varshney, D., Yari, M., and Murphy, R. (2020). “FloodNet: A High Resolution Aerial Imagery Dataset for Post Flood Scene Understanding.”.
Rajchl, M., et al. (2017). “DeepCut: Object Segmentation From Bounding Box Annotations Using Convolutional Neural Networks.” IEEE Transactions on Medical Imaging, 36(2), 674–683.
Saleh, F., Aliakbarian, M. S., Salzmann, M., Petersson, L., Gould, S., and Alvarez, J. M. (2016). “Built-in Foreground/Background Prior for Weakly-Supervised Semantic Segmentation.” Lecture Notes in Computer Science, 413–432.
Selvaraju, R. R., Cogswell, M., Das, A., Vedantam, R., Parikh, D., and Batra, D. (2017). “Grad-CAM: Visual Explanations from Deep Networks via Gradient-Based Localization.” 2017 IEEE International Conference on Computer Vision (ICCV), IEEE, 618–626.
Zhu, X., Liang, J., and Hauptmann, A. (2020). “MSNet: A Multilevel Instance Segmentation Network for Natural Disaster Damage Assessment in Aerial Videos.”, 2023–2032.
Information & Authors
Information
Published In
History
Published online: Mar 18, 2024
ASCE Technical Topics:
- Architectural engineering
- Architecture
- Benefit cost ratios
- Buildings
- Business management
- Channels (waterway)
- Computer models
- Computer vision and image processing
- Education
- Engineering fundamentals
- Financial management
- Hydraulic engineering
- Hydraulic structures
- Methodology (by type)
- Models (by type)
- Practice and Profession
- Structural engineering
- Structures (by type)
- Training
- Water and water resources
- Waterways
Authors
Metrics & Citations
Metrics
Citations
Download citation
If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download.