Spatial-Resolution Independent Object Detection Framework for Aerial Imagery

Samanta, Sidharth; Panda, Mrutyunjaya; Ramasubbareddy, Somula; Sankar, Shrinivasan; Burgos, Daniel

doi:10.32604/cmc.2021.014406

Cited by 3 publications

(4 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Here, the metrics are calculated based on the is_iceberg attribute of the given JSON dataset file. The formula used to calculate precision, recall, and the F1 score is provided in Equations ( 12)- (14).…”

Section: Performance Analysis Of Resultsmentioning

confidence: 99%

“…GAN is composed of two models: generative and discriminative models. The generator model uses the network, which generates images by adding random objects [14] and produces an input image that is similar to the actual image. The discriminator attempts to classify whether the generated image the generator generates looks like the actual input image.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Aggregate Channel Features and Fast Regions CNN Approach for Classification of Ship and Iceberg

Sethu Ramasubiramanian,

Sivasubramaniyan,

Peer Mohamed

2023

Applied Sciences

View full text Add to dashboard Cite

Detection and classification of icebergs and ships in synthetic aperture radar (SAR) images play a vital role in marine surveillance systems even though available adaptive threshold methods give satisfying results on detection and classification for ships and icebergs, including techniques of convolutional neural networks (CNNs), but need more accuracy and precision. An efficient and accurate method was developed to detect and classify the ship and icebergs. Hence, the research method proposed locating and classifying both ships and icebergs in a given SAR image with the help of deep learning (DL) and non-DL methods. A non-DL method utilized here was the aggregate channel features (ACF) detector, which extracts region proposals from huge SAR images. The DL object detector called fast regions CNN (FRCNN) detects objects accurately from the result of ACF since the ACF method avoids unwanted regions. The novelty of this study was that ACF-FRCNN concentrates only on accurately classifying ships and icebergs. The proposed ACF-FRCNN method gave a better performance in terms of loss (18.32%), accuracy (96.34%), recall (98.32%), precision (95.97%), and the F1 score (97.13%). Compared to other conventional methods, the combined effect of ACF and FRCNN increased the speed and quality of the detection of ships and icebergs. Thus, the ACF-FRCNN method is considered a novel method for over 75 × 75 resolution ship and iceberg SAR images.

show abstract

Section: Performance Analysis Of Resultsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Aggregate Channel Features and Fast Regions CNN Approach for Classification of Ship and Iceberg

Sethu Ramasubiramanian,

Sivasubramaniyan,

Peer Mohamed

2023

Applied Sciences

View full text Add to dashboard Cite

show abstract

“…GAN composite two models which are Generative and Discriminative model. The generator model uses the network which generates images by adding random [14] and produces an input image like a real input image and discriminator attempts to classify whether the generated image is really it looks like real input image or not. In [15], the super resolution network was integrated with the cycle model based on GAN residuals.…”

Section: Existing Deep Learning Methods For Small Object Detectionmentioning

confidence: 99%

Aggregate Channel Features and Fast Rcnn for Ship, Iceberg Classification

M.S.,

Suresh,

Fathimal

2023

Preprint

View full text Add to dashboard Cite

Detection and classification of Iceberg and ship in synthetic aperture radar (SAR) image plays a vital role in marine surveillance systems. Even though available adaptive thresholding methods give satisfying results on this problem, still those methods face complexity on classification of ship and icebergs. The research proposed aims to locate and classify both ship and iceberg in a given SAR image with the help of deep learning and non-deep learning methods. The non-deep learning method, Aggregate Channel Features Detector (ACF), was used to extract region proposals from the huge SAR images. The deep learning object detector, Fast R-CNN detects objects accurately from the result of ACF by avoiding unwanted regions and concentrates only on accurate classification of ship and iceberg. The proposed method gives better performance in terms of accuracy, recall, precision and F1 score rather than other conventional methods and reduces the complexity of detector.

show abstract

“…The onestage method uses the anchor method to decide whether each grid is on the center point of the object, and finally perform regression and classification of the detection of the object without using the candidate boxes. Generally speaking, the two-stage method has better detection accuracy than the onestage method [20], while the one-stage method weighted more than two-stage method in the aspect of detection speed. Generally, both one-stage and two-stage methods can achieve good results in object detection.…”

Section: Related Work 21 Object Detectionmentioning

confidence: 99%

Few-Shot Object Detection Based on the Transformer and High-Resolution Network

Zhang¹,

Pu²,

Li³

et al. 2023

Computers, Materials &Amp; Continua

View full text Add to dashboard Cite

Now object detection based on deep learning tries different strategies. It uses fewer data training networks to achieve the effect of large dataset training. However, the existing methods usually do not achieve the balance between network parameters and training data. It makes the information provided by a small amount of picture data insufficient to optimize model parameters, resulting in unsatisfactory detection results. To improve the accuracy of few shot object detection, this paper proposes a network based on the transformer and high-resolution feature extraction (THR). High-resolution feature extraction maintains the resolution representation of the image. Channels and spatial attention are used to make the network focus on features that are more useful to the object. In addition, the recently popular transformer is used to fuse the features of the existing object. This compensates for the previous network failure by making full use of existing object features. Experiments on the Pascal VOC and MS-COCO datasets prove that the THR network has achieved better results than previous mainstream few shot object detection.

show abstract

Spatial-Resolution Independent Object Detection Framework for Aerial Imagery

Cited by 3 publications

References 18 publications

Aggregate Channel Features and Fast Regions CNN Approach for Classification of Ship and Iceberg

Aggregate Channel Features and Fast Regions CNN Approach for Classification of Ship and Iceberg

Aggregate Channel Features and Fast Rcnn for Ship, Iceberg Classification

Few-Shot Object Detection Based on the Transformer and High-Resolution Network

Contact Info

Product

Resources

About