Intensive Positioning Network for Remote Sensing Image Captioning

Wang, Shengsheng; Chen, JiaWei; Wang, Guangyao

doi:10.1007/978-3-030-02698-1_49

Cited by 7 publications

(8 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In Attention mechanism helps the encoderdecoder algorithm in concentrating on important parts of an image so the generated caption will be capturing main aspect of the image. This approach for caption generation is followed in [15,49,50,[61][62][63][64].…”

Section: Encoder-decoder Approachmentioning

confidence: 99%

“…To avoid information loss while captioning an aerial image due to higher pixels and smaller target size, Intensive Positioning Network (IPN) is proposed in [61]. The algorithmic view of this model is shown in Figure 5.…”

Section: Encoder-decoder Approachmentioning

confidence: 99%

“…Figure 5. Algorithmic view of model proposed in [61] Special characteristics of aerial images such as scale ambiguity, category ambiguity, and rotation ambiguity are considered for caption generation in [49]. The proposed approach is based on encoderdecoder framework [27] and shown in Figure 6.…”

Section: Encoder-decoder Approachmentioning

confidence: 99%

See 2 more Smart Citations

Automatic Caption Generation for Aerial Images: A Survey

Mondhe

Satone

Kharate

2023

IJEEI

View full text Add to dashboard Cite

Aerial images have attracted attention from researcher community since long time. Generating a caption for an aerial image describing its content in comprehensive way is less studied but important task as it has applications in agriculture, defence, disaster management and many more areas. Though different approaches were followed for natural image caption generation, generating a caption for aerial image remains a challenging task due to its special nature. Use of emerging techniques from Artificial Intelligence (AI) and Natural Language Processing (NLP) domains have resulted in generation of accepted quality captions for aerial images. However, lot needs to be done to fully utilize potential of aerial image caption generation task. This paper presents detail survey of the various approaches followed by researchers for aerial image caption generation task. The datasets available for experimentation, criteria used for performance evaluation and future directions are also discussed.

show abstract

Section: Encoder-decoder Approachmentioning

confidence: 99%

Section: Encoder-decoder Approachmentioning

confidence: 99%

See 1 more Smart Citation

Automatic Caption Generation for Aerial Images: A Survey

Mondhe

Satone

Kharate

2023

IJEEI

View full text Add to dashboard Cite

show abstract

“…RSI captioning has been an interesting topic in RSI processing combined with the natural language processing [1], [11]- [13], [20], [28], [29]. According to the way of generating sentences, the main methods can be divided into three categories: templatebased methods, retrieval-based methods, and RNN-based methods.…”

Section: A Rsi Captioningmentioning

confidence: 99%

“…Considering the scale ambiguity, category ambiguity, and rotation ambiguity of RSI, Lu et al [13] explore the RSI captioning task based on both the hand-crafted features and deep features. A captioning method is proposed by Wang et al [28], which can provide more accurate object location via adding additional location and category tags to images. The information from the fully connected layer of CNN is treated as attribute information to improve the performance of captioning method based on convolutional features [29].…”

Section: A Rsi Captioningmentioning

confidence: 99%

Retrieval Topic Recurrent Memory Network for Remote Sensing Image Captioning

Wang

Zheng

2020

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

Remote sensing image (RSI) captioning aims to generate sentences to describe the content of RSIs. Generally, five sentences are used to describe the RSI in caption datasets. Every sentence can just focus on part of images' contents due to the different attention parts of annotation persons. One annotated sentence may be ambiguous compared with other four sentences. However, previous methods, treating five sentences separately, may generate an ambiguous sentence. In order to consider five sentences together, a collection of words, which named topic words contained common information among five sentences, is jointly incorporated into a captioning model to generate a determinate sentence that covers common contents in RSIs. Instead of employing a naive recurrent neural network, a memory network in which topic words can be naturally included as memory cells is introduced to generate sentences. A novel retrieval topic recurrent memory network is proposed to utilize the topic words. First, a topic repository is built to record the topic words in training datasets. Then, the retrieval strategy is exploited to obtain the topic words for a test image from topic repository. Finally, the retrieved topic words are incorporated into a recurrent memory network to guide the sentence generation. In addition to getting topics through retrieval, the topic words of test images can also be edited manually. The proposed method sheds light on controllability of caption generation. Experiments are conducted on two caption datasets to evaluate the proposed method.

show abstract

Remote Sensing Image Captioning via Multilevel Attention-Based Visual Question Answering

Murali

Shanthi

2022

Advances in Intelligent Systems and Computing

View full text Add to dashboard Cite

Intensive Positioning Network for Remote Sensing Image Captioning

Cited by 7 publications

References 8 publications

Automatic Caption Generation for Aerial Images: A Survey

Automatic Caption Generation for Aerial Images: A Survey

Retrieval Topic Recurrent Memory Network for Remote Sensing Image Captioning

Remote Sensing Image Captioning via Multilevel Attention-Based Visual Question Answering

Contact Info

Product

Resources

About