Object Detection in Optical Remote Sensing Images Based on Weakly Supervised Learning and High-Level Feature Learning

By Junwei Han, Dingwen Zhang, Gong Cheng, Lei Guo, Jinchang Ren. 2015

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

1. Gong Cheng
2. Junwei Han
3. Liangpei Zhang
4. Gui-Song Xia
5. Xiaoqiang Lu
6. Ke Li
7. Xiwen Yao
8. Bo Du

Citations in the corpus

• EarthGPT: A Universal Multimodal Large Language Model for Multisensor Image Comprehension in Remote Sensing Domain (2024, methodology): “When directly using MMRS-1M for fine-tuning, EarthGPT achieves a Top-1 accuracy of 67.35%.”
• Using convolutional features and a sparse autoencoder for land-use scene classification (2016, methodology): “Han et al. (2015) proposed a weakly supervised learning (WSL) framework for object detection from optical images.”
• Novel segmented stacked autoencoder for effective dimensionality reduction and feature extraction in hyperspectral imaging (2016, background): “Future work will explore the combination of most state-of-the-art techniques including 2D SSA [22], adaptive sparse representation [23, 27], Hybrid and sampling-based clustering ensemble [24], weakly supervised learning [25], tensor rank selection [26], gradient and subspace processing [28] and sali”
• Object detection in VHR optical remote sensing images via learning rotation-invariant HOG feature (2016, methodology): “We therefore adopt a simple and widely-used non-maximum suppression (NMS) strategy as the work of [1, 3, 4, 6, 7, 9] to eliminate the repeated detections.”
• Building Instance Classification Using Street View Images (2018, background): “By exploiting deep Boltzmann machine, a novel weakly supervised learning approach for object detection in remote sensing images is introduced (Han et al., 2015).”
• Ship Rotated Bounding Box Space for Ship Extraction From High-Resolution Optical Satellite Images With Complex Backgrounds (2016, methodology): “[9] detected typical objects in remote sensing images based on square sliding windows in varying sizes by a detector trained using a new weakly supervised learning framework.”
• Accurate Object Localization in Remote Sensing Images Based on Convolutional Neural Networks (2017, background): “feature transform (SIFT) [1], histogram of oriented gradients (HOG) [2], and Saliency [3], [4].”
• High-resolution remote-sensing imagery retrieval using sparse features by auto-encoder (2015, unknown): “Some other similar works can be found in Cheng et al. (2015) and Han et al. (2014, 2015).”
• Semantic Segmentation based on Stacked Discriminative Autoencoders and Context-Constrained Weakly Supervised Learning (2015, background): “…such as high-level features learned by recent developed deep learning methods, may offer a promising venue to empower the model to handle intrinsically diverse and complex visual classes in natural images [8] and also remote sensing images [9], and hence to improve segmentation performance.”
• Rotation-Insensitive and Context-Augmented Object Detection in Remote Sensing Images (2018, methodology): “The model has been applied with success in a variety of natural scene images and remote sensing images [21], [48], [49].”
• Exploring Models and Data for Remote Sensing Image Caption Generation (2017, background): “W ITH the development of remote sensing technology, remote sensing images have become available and attracted extensive attention in numerous applications [1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15].”
• Incremental Detection of Remote Sensing Objects With Feature Pyramid and Knowledge Distillation (2020, methodology): “With the aid of studies of weakly supervised object detection in remote sensing images [46], [47], a more feasible way is to employ class-incremental detection with image-level annotations, which is also a possible topic in our feature work.”
• Learning a Rotation Invariant Detector with Rotatable Bounding Box (2017, methodology): “As another effective feature extraction tool, deep Boltzmann machine is also used as feature extractor in geo-spatial object detection problems [11, 7].”