Name
Abstract | ||
|---|---|---|
Deep learning algorithms for connectomics rely upon localized classification, rather than overall morphology. This leads to a high incidence of erroneously merged objects. Humans, by contrast, can easily detect such errors by acquiring intuition for the correct morphology of objects. Biological neurons have complicated and variable shapes, which are challenging to learn, and merge errors take a multitude of different forms. We present an algorithm, MergeNet, that shows 3D ConvNets can, in fact, detect merge errors from high-level neuronal morphology. MergeNet follows unsupervised training and operates across datasets. We demonstrate the performance of MergeNet both on a variety of connectomics data and on a dataset created from merged MNIST images. |
Year | Venue | Field |
|---|---|---|
2017 | arXiv: Computer Vision and Pattern Recognition | Connectomics,MNIST database,Pattern recognition,Computer science,Intuition,Error detection and correction,Artificial intelligence,Deep learning,Merge (version control),Machine learning |
DocType | Volume | Citations |
Journal | abs/1705.10882 | 0 |
PageRank | References | Authors |
0.34 | 15 | 8 |
Name | Order | Citations | PageRank |
|---|---|---|---|
| David Rolnick | 1 | 65 | 10.53 |
| Yaron Meirovitch | 2 | 25 | 3.33 |
| Toufiq Parag | 3 | 52 | 7.18 |
| Hanspeter Pfister | 4 | 5933 | 340.59 |
| Viren Jain | 5 | 294 | 36.11 |
| Jeff W. Lichtman | 6 | 134 | 12.41 |
| E. Boyden, III | 7 | 31 | 6.37 |
| Nir Shavit | 8 | 3780 | 244.84 |