ADEPT: advanced dense prediction

Semantic analysis of natural images at the pixel level

The project ADEPT studies dense semantic analysis in large images of natural scenes. Our goal is to reduce obstacles towards easy deployment in exciting real-world applications such as autonomous cars, road safety inspection or automated warehouses.

orsic21pr bevandic19gcpr saric20cvpr

Topics

We are especially interested in the following problems:

Methodology

We study convolutional models for dense prediction based on checkpointed DenseNet backbones, lightweight ladder-style upsampling [1] and pyramidal fusion [2]. We predict the future semantic content in video based on feature-to-motion-and-feature forecasting [3]. We apply multi-domain loss expressed as negative log likelihood of aggregated probability [4]. We address open-set recognition throughout learning with noisy and artificial negative samples [5].

References:

[1] ladder-DenseNets, T-ITS 2020.
[2] pyramidal fusion, CVPR 2019, PR 2021.
[3] F2MF forecasting, CVPR 2020.
[4] multi-domain training, NLL+ loss, Arxiv.
[5] open-set recognition with noisy negatives, GCPR 2019.

Time frame

Start date: 1st February 2021.

Duration: 48 months

The project has been fully funded by the Croatian science foundation