Zoeken
Zoeken kan via de modus 'eenvoudig zoeken' (één veld) of uitgebreid via 'geavanceerd zoeken' (meerdere velden). Zo kan je bv. zoeken op een combinatie van een auteursnaam (auteur), een jaartal (jaar) en een documenttype.
Boekenmand
Nuttige resultaten kan je aanvinken en toevoegen aan een mandje. De inhoud hiervan kan je exporteren of afdrukken (naar bv. PDF).
RSS
Op de hoogte blijven van nieuw toegevoegde publicaties binnen uw interessegebied? Dit kan door een RSS-feed (?) te maken van jouw zoekopdracht.
nieuwe zoekopdracht
Underwater image recognition detector using deep ConvNet
|
| Beschikbaar in | Auteurs |
|
Documenttype: Congresbijdrage
|
| Auteurs | | Top |
- Lakshmi, M.D.
- Santhanam, S.M.
|
|
|
| Abstract |
Underwater navigation and intelligent object recognition require robust machine learning algorithms to operate in turbid water. Modern life created the man-made pollution in oceans, rivers, and lakes, which contaminate our water resources. Despite environmental regulations solid waste in the form of trash, litter and garbage are thrown directly into sea spoiling the existence of underwater living creatures. The underwater vehicle can be used for survey purposes. The key challenge of underwater image-based localization comes from the unstructured nature of the seabed terrain. So, there is a need for robust detection of the features in such environments is essential. Hence, this paper proposes the automated underwater image recognition detector for submersible imagery. We train a Convolutional neural Network (ConvNet) to classify input 64 × 64 images and considered the classifier as an object feature detector. The features of the image from underwater-bed can be extracted and forward into a network. The output of the three-layer ConvNet with deeply connected network results in a probability distribution over N classes. A Stochastic gradient descent with ADAM optimizer uses the squared gradients to scale the learning rate and reduces the difference between the actual and predicted output. The evaluations are done on the precision, recall, F-Score, macro and weighted Average accuracy for both the detectors. It is observed that our proposed network, achieved an overall accuracy of 93.9 % for correct detections with a binary detector and 90.1% with a multiclass detector compared to existing detectors. |
IMIS is ontwikkeld en wordt gehost door het VLIZ.
