Course paper / final thesis

Studying pre-training datasets for improving classification and detection in sidescan sonar images

Starting date


Duration of contract

initially limited to 6 months

Type of employment

Full-time (part-time possible)

Since sidescan sonar image data is scarce, transfer-learning of deep learning models is typically applied. The models are first pre-trained on a large dataset, e.g. ImageNet for classification or MS COCO for detection, and afterwards fine-tuned on the sonar data. Those pre-training datasets consist of natural RGB images. Sonar images, however, are grayscaled intensity images. Thus, learned features which depend on color information are useless. Using a more suited pre-training dataset could improve the classification and detection performance of deep learning models on sidescan sonar images.

In general, recent work has shown that the pre-training dataset has a strong influence when fine-tuning deep learning models, especially when the domain gap (the difference between data in the pre-training and fine-tuning dataset) is large. This is the case for sonar images and standard pre-training dataset like ImageNet or MS COCO. A survey on alternative and more suited pre-training datasets should be carried out in this master thesis. Both computer vision tasks classification and detection should be considered.

Your tasks:

  • literature review on transfer-learning and the influence of the pre-training
  • search for alternative datasets with smaller domain gap (e.g. from medical or radar applications)
  • analyze classification and detection performance for selected datasets
  • compare different deep learning architectures

Your qualifications:

  • studying electronics engineering, computer science or equivalent
  • knowledge in the fields of deep learning and computer vision
  • knowledge in underwater acoustics and sonar signal processing beneficial
  • programming skills in Python required

Your benefits:

Look forward to a fulfilling job with an employer who appreciates your commitment and supports your personal and professional development. Our unique infrastructure offers you a working environment in which you have unparalleled scope to develop your creative ideas and accomplish your professional objectives. Our human resources policy places great value on a healthy family and work-life-balance as well as equal opportunities for persons of all genders (f/m/x). Individuals with disabilities will be given preferential consideration in the event their qualifications are equivalent to those of other candidates.

  • Apply online now
  • You can send this job advertisement via e-mail and complete your application on a personal computer or laptop.

    We need your digital application documents (PDF). The document upload function is not supported by all mobile devices. Please complete your application on a PC/laptop.

    Complete application on PC

Technical contact

Yannik Steiniger
Institute for the Protection of Maritime Infrastructures

Phone: +49 471 924199-53

Send message

Dr. Jannis Ulrich Stoppe
Institute for the Protection of Maritime Infrastructures

Phone: +49 471 924199-43

Send message

Vacancy 78743

HR department Göttingen

Send message

DLR site Bremerhaven

To location

DLR Institute for the Protection of Maritime Infrastructures

To institute