Abstract | ||
---|---|---|
We present an approach that leverages multiple datasets possibly annotated using different classes to improve the semantic segmentation accuracy on each individual dataset. We propose a new selective loss function that can be integrated into deep networks to exploit training data coming from multiple datasets with possibly different tasks (e.g., different label-sets). We show how the gradient-reversal approach for domain adaptation can be used in this setup. Thorought experiments on semantic segmentation applications show the relevance of our approach. |
Year | DOI | Venue |
---|---|---|
2016 | 10.1007/978-3-319-49055-7_30 | Lecture Notes in Computer Science |
Keywords | Field | DocType |
Deep learning,Convolutional neural networks,Semantic segmentation,Domain adaptation,Multi-ask learning | Scale-space segmentation,Multi-task learning,Pattern recognition,Domain adaptation,Segmentation,Computer science,Convolutional neural network,Exploit,Artificial intelligence,Deep learning,Semantic computing,Machine learning | Conference |
Volume | ISSN | Citations |
10029 | 0302-9743 | 0 |
PageRank | References | Authors |
0.34 | 23 | 6 |
Name | Order | Citations | PageRank |
---|---|---|---|
Damien Fourure | 1 | 8 | 1.62 |
Rémi Emonet | 2 | 61 | 7.60 |
Élisa Fromont | 3 | 192 | 25.51 |
Damien Muselet | 4 | 167 | 16.88 |
alain tremeau | 5 | 230 | 34.42 |
Christian Wolf | 6 | 1027 | 54.93 |