WEAKLY SUPERVISED DEEP CONVOLUTIONAL NETWORKS FOR FINE-GRAINED OBJECT RECOGNITION IN MULTISPECTRAL IMAGES


Creative Commons License

Aygunes B., AKSOY S., CİNBİŞ R. G.

IEEE International Geoscience and Remote Sensing Symposium (IGARSS), Yokohama, Japonya, 28 Temmuz - 02 Ağustos 2019, ss.1478-1481 identifier identifier

  • Cilt numarası:
  • Doi Numarası: 10.1109/igarss.2019.8899170
  • Basıldığı Şehir: Yokohama
  • Basıldığı Ülke: Japonya
  • Sayfa Sayıları: ss.1478-1481

Özet

The challenging task of training object detectors for fine-grained classification faces additional difficulties when there are registration errors between the image data and the ground truth. We propose a weakly supervised learning methodology for the classification of 40 types of trees by using fixed-sized multispectral images with a class label but with no exact knowledge of the object location. Our approach consists of an end-to-end trainable convolutional neural network with separate branches for learning class-specific and location-specific scoring of image regions. Comparative experiments show that the proposed method simultaneously learns to detect and classify the objects of interest with high accuracy.