ADEPT: advanced dense prediction

Publications

Conferences

• Matej Grcic, Josip Saric, Sinisa Segvic: On Advantages of Mask-level Recognition for Outlier-aware Segmentation. CVPR Workshops 2023: 2937-2947. (cvf).
• Nishant Kumar, Sinisa Segvic, Abouzar Eslami, Stefan Gumhold: Normalizing Flow based Feature Synthesis for Outlier-Aware Object Detection. CVPR 2023: 5156-5165. (cvf).
• Petra Bevandic, Sinisa Segvic. Automatic universal taxonomies for multi-domain semantic segmentation. BMVC 2022: 63. (bmvc).
• Matej Grcic, Petra Bevandic, Sinisa Segvic: DenseHybrid: Hybrid Anomaly Detection for Dense Open-Set Recognition. ECCV (25) 2022: 500-517. (ecva).
• Petra Bevandić, Marin Oršić, Ivan Grubišić, Josip Šarić, Siniša Šegvić. Multi-domain semantic segmentation with overlapping labels. WACV 2022. (cvf).
• Ivan Grubisic, Marin Orsic, Sinisa Segvic. A baseline for semi-supervised learning of efficient semantic segmentation models. MVA 2021. (pdf)
• Matej Grcić, Ivan Grubišić, Siniša Šegvić. Densely connected normalizing flows. NeurIPS 2021. (neurips, arxiv)

Journals

• Petra Bevandic, Ivan Kreso, Marin Orsic, Sinisa Segvic: Dense open-set recognition based on training with noisy negative images. Image Vis. Comput. 124: 104490 (2022) (ScienceDirect, arxiv).
• Josip Saric, Marin Orsic, Sinisa Segvic: Panoptic SwiftNet: Pyramidal Fusion for Real-Time Panoptic Segmentation. Remote. Sens. 15(8): 1968 (2023) (doi).
• Ivan Grubisic, Marin Orsic, Sinisa Segvic: Revisiting Consistency for Semi-Supervised Semantic Segmentation. Sensors 23(2): 940 (2023) (doi).
• Iva Sović, Josip Šarić, Siniša Šegvić. Dense Semantic Forecasting with Multi-Level Feature Warping. Appl. Sci. 2023, 13(1), 400. (doi).
• Josip Šarić, Sacha Vražić, Siniša Šegvić. Dense Semantic Forecasting in Video by Joint Regression of Features and Feature Motion. IEEE TNNLS 2021. (ieee, arxiv).

Other recent work

• Marin Oršić, Siniša Šegvić. Efficient semantic segmentation with pyramidal fusion. Pattern Recognit. 110: 107611, 2021 (elsevier).
• Ivan Krešo, Josip Krapac, Siniša Šegvić. Efficient Ladder-Style DenseNets for Semantic Segmentation of Large Images. IEEE Trans. Intell. Transp. Syst. 22(8): 4951-4961, 2021 (ieee)

Theses

• Petra Bevandić: Open-set segmentation of images by means of negative examples (pdf). UniZg-FER 2023.
• Josip Šarić: Združeno prognoziranje značajki i njihova pomaka za predviđanje semantičke budućnosti u videu (pdf). UniZg-FER 2022.
• Marin Oršić: Efficient semantic image segmentation using pyramidal fusion (pdf). UniZg-FER 2021.
• Ivan Krešo: Convolutional architecture for efficient semantic segmentation of large images (pdf). UniZg-FER 2021.